RU2193825C2 - Method and device for processing signals to find coordinates of objects displayed as sequence of television images - Google Patents

Abstract

FIELD: computer-aided systems for controlling street traffic and landing maneuvers in airports, robotics, etc. SUBSTANCE: proposed method and device intended for processing images in miscellaneous systems for viewing and recording behavior of various objects and for functioning as subsystems of higher-level interpreting systems for locating, segmenting, and viewing moving objects and also for automatic determination of their parameters, locating their coordinates, and steadily holding object image in field of vision of observation system video camera or in center of tracking window upon reception of both contrast and low-contrast images of fixed and motionless objects and background of terrain from video camera involve adaptive and simultaneous use of several complementary pixel classification channels, image shaping, and location of object coordinates including compensation of destabilizing factors. Proposed alternative methods are implemented by respective alternative devices. EFFECT: enlarged functional capabilities. 56 cl, 18 dwg

Description

 The invention relates to the field of processing television images in various machine vision applications, for example, in monitoring and control tasks, in particular to methods for determining the position of an object from a sequence of television images, and can be used in automated traffic control systems for monitoring and documenting take-off and landing maneuvers at airports, in robotics to ensure coordinated interaction of the machine with other machines or production objects and a more general approach can serve as a subsystem for any systems of higher level interpretation, through which detected, there may be segmented and still and moving objects and their parameters are determined automatically.

 Known methods and devices for monitoring objects, on the basis of which it is possible to implement the simplest systems of the first generation, are characterized by using one of the basic approaches for classifying image elements into elements of the object and background (generating a binary image of the object) and determining the coordinates of the object by determining the coordinates of the center of gravity of the binarized image (the so-called centroid methods and devices) or by determining the position of the object based on the search for the extremum of the measure c similarities / dissimilarities (in particular, the difference correlation function) of the current and reference images (correlation-extreme methods and devices).

 Such systems make it possible to automatically determine the coordinates of an object with sufficient accuracy and ensure stable retention of an image of an object near the center of the field of view only in those cases when it is highly contrasted with respect to the background and (or) when the image of the object changes little during observation, and the effect of other destabilizing the image of factors is absent or minimal. These limitations result in a relatively narrow range of observation conditions for objects in which these systems are applicable.

 For many real-world applications, the visible images of the object are either low contrast with respect to the background, or (with medium and high contrast) observed against the contrasting background of the area, and due to the movement of the object or the observation system, the image of the object changes radically during the observation time. The situation is often complicated by the fact that the object can be partially or completely temporarily shielded in the image of the scene by the objects of the background, and destabilizing factors act on the video camera, leading to uncontrolled horizontal and vertical movements and rotation of the field of view.

 A known method for determining the coordinates of an object (see copyright certificate 1562980 USSR, MKI5 N 04 N 7/18, Device for determining the coordinates of an object / B.A. Alpatov, P.A.Bakut, I.E. Vornovitsky, A.A. Selyaev, A.I. Stepashkin, S. Yu. Khludov (USSR), - 4373608 / 24-09; Application. 01.02.88; Publish. 07.05.90, Bull. 17), consisting of the following set of actions.

 At the same time, the current frame is recorded and the previous frame is processed.

 From the first received frame form the target standard h (i, j) of the MXM format by cutting out the central part of the image.

 Starting from the 3rd received frame, the previous recorded frame is processed as follows.

где g(i+ν, j+μ) - элементы текущего обрабатываемого изображения размерностью NxN точек,
h(i,j) - элементы эталонного изображения размерностью МхМ точек,
Z(i, j) - элементы весовой функции размерностью МхМ точек, причем Z(i,j) = 1, если точка изображения принадлежит объекту, Z(i,j) = 0 - иначе;
ν,μ - параметры сдвига изображения, которые изменяются от 0 до ±(N-M+1).The weighted difference correlation function F (ν, μ) is calculated

where g (i + ν, j + μ) are the elements of the current processed image of dimension NxN points,
h (i, j) - elements of the reference image with a dimension of MxM points,
Z (i, j) - elements of the weight function with the dimension МХМ points, moreover, Z (i, j) = 1 if the image point belongs to the object, Z (i, j) = 0 otherwise;
ν, μ - image shift parameters that vary from 0 to ± (N-M + 1).

 To reduce the number of computational operations, the calculation of the difference correlation function F (ν, μ) at a point with coordinates ν, μ is performed until the current value of this function remains less than some threshold function R (k), which depends on the number k of elementary differences calculated at this point in time.

The coordinates ν ^* , μ ^{* of the} object in the image are determined based on the minimization of the difference correlation function F (ν, μ).
Produce the current low-frequency (LF) filtering (smoothing) of the image
V _n (i, j) = (1-β) g _n (i + ν $\genfrac{}{}{0ex}{}{\text{*}}{\text{n}}$ , j + μ $\genfrac{}{}{0ex}{}{\text{*}}{\text{n}}$ ) + βV _n-1 (i, j), (2)
where V _n (i, j) is the value of the smoothed image in the nth frame,
β is the smoothing parameter, 0 <β <1.

и определения перемещения объекта за последние L кадров. Если найденное перемещение объекта составило более одного дискрета, то производится перезапись эталона
h(i,j)=V_n(i,j). (4)
Если перемещение объекта за последние L кадров не превысило одного дискрета, то перезапись эталона выполняют при условии F(ν^*,μ^*)>Порог, иначе эталон сохраняют.The conditions for updating the reference image are determined by calculating the duration of the transient process L of the smoothing filter (2)

and determining an object’s movement over the last L frames. If the found displacement of the object was more than one discrete, then the standard is overwritten
h (i, j) = V _n (i, j). (4)
If the movement of an object over the last L frames does not exceed one discrete, then rewriting of the standard is performed under the condition F (ν ^* , μ ^* )> Threshold, otherwise the standard is saved.

 Starting with L + 1 frames, the set of Q points belonging to the object is selected on the smoothed image. The weight function Z (i, j) in the first L + 1 frames is taken equal to 1.

 The selection of the set of Q points belonging to the object is produced by threshold processing of the smoothed image.

 Many Q points belonging to the object are used to form a new weight function Z (i, j) of the reference image, which is changed simultaneously with the change of the standard.

 A sign of the method that coincides with the sign of the proposed technical solution is the calculation of a weighted measure of the dissimilarity of the image signals in the current analysis window with the reference image of the object in the form of a difference correlation function F (ν, μ) in accordance with expression (1).

The reasons hindering the achievement of the required technical result when using this method are to determine the coordinates ν ^* , μ ^{* of the} image of the object based on minimizing the difference correlation function F (ν, μ) using the weight function Z (ij) in the first L + 1 frames, equal to 1, which makes it impossible to correctly determine the coordinates of the object in conditions when the contrast of the object is lower than the contrast of the elements of the background surrounding it.

 The reasons for the inoperability of the method in the described situations are illustrated by figa, 1b, 1c, 1d.

 In FIG. Figure 1 shows the formation of a standard from the first received frame (Fig. 1a) by cutting out its central part (Fig. 1b). At this stage, the standard is not yet protected from extraneous background elements. Typical for this situation, the distribution of the brightness of the image elements of the generated standard and the form of the difference correlation function F (ν, μ) are illustrated in FIG.

 As can be seen from FIG. With a contrasting motley background, the false minimum is smaller than the minimum corresponding to the position of the object. Thus, the assessment of the coordinates of the object in accordance with this method leads to an erroneous determination of coordinates and disruption of tracking.

 The reasons that impede the achievement of the required technical result can also include a combination of updating the standard and calculating the difference correlation function of the current and reference images, which leads to tracking failures when even relatively small areas of the object (10-20% of the area) overlap with high-contrast background elements.

 The reasons for the inoperability of the method in such situations are illustrated fig.1d, 1e.

 In the conditions under consideration, we can proceed from the fact that during the initial stage of tracking, the background along which the object was moving was such that the reference image of the object was formed correctly, that is, there are no background elements in the standard (Fig. 1d).

 When moving an object behind contrasting objects of the background, as shown in FIG. 1f, the minimum value of the difference correlation function increases sharply and, in accordance with the sequence of operations of the method, this causes a rewriting of the standard. As a result, the contrasting background element falls into the standard, and then, in the same way as described above, erroneous determination of the coordinates of the object and disruption of tracking occur.

 In addition, in the presence of unpredictable movements of the field of view during the rewriting of the pattern, background elements fall into the video window of the pattern of the object, which, as described above, also often leads to disruption of tracking.

 To ensure the operability of systems for determining coordinates and observing objects in such an extended range of conditions, modern methods and devices (conditionally methods and devices of the second generation) are developing in the direction of adaptive and simultaneous use of several mutually complementary channels for classifying image elements, image formation and determining the coordinates of an object, accounting and compensation of destabilizing factors.

 Known methods and devices for determining the coordinates of objects by the time sequence of images do not have the necessary universality for functioning in a wide range of application conditions, namely, in conditions of high and low contrast images of the object and the surrounding background of the terrain, with changes in the size and visible image of the object as a result of its maneuvers, and as a result of its movement in the area, changes or uneven illumination of the object and the terrain, the screening of the object by other objects and the portions or areas, movement of the support monitoring and determining the parameters of the objects, absence or incomplete (insufficient) stabilizing video cameras field of view and others.

 A known method of detecting and determining the coordinates of an object in an image (see copyright certificate 1737755 USSR, MKI 5 H 04 N 7/18, G 06 K 9/36, Device for detecting and determining the coordinates of an object in an image / B. A. Alpatov, E. T.Libiyainen, S.Yu. Khludov (USSR) Ryazan Radio Engineering Institute (USSR), - 4819727/09, Declaration 04/27/90; Publish. 05/30/92, Bul. 20), the implementation of which requires the following set of actions.

At the first stage, in the first frame, only the television image is recorded in the memory. During the recording of the second frame, the initial data arrays are formed: G ₀ - background reference image G ₀ = L ₀ - current image. (The current image is taken as the background reference image. It is assumed that there is no object in the image L _0. )
B0 is the initial array of elements of the object / background classification function,
B0: b0ij = -32.

 Starting from the 3rd received frame, write to the memory of the current frame.

где 1_nij - элементы текущего обрабатываемого изображения L_n,

- элементы прогнозируемого изображения фона,
D=1,

если b_(n-1)ij ≤ 0,
Δ1 - прогнозируемое минимальное отличие яркости объекта от яркости фона.At the second stage (object detection stage), for each element lnij of the observed image, elements bnij of the classification function are determined

where 1 _nij are the elements of the current processed image L _n ,

- elements of the predicted background image,
D = 1

if b _{(n-1) ij} ≤ 0,
Δ1 is the predicted minimum difference in the brightness of the object from the brightness of the background.

Make a decision on whether the image element with coordinates i, j belongs to the object if the resulting value is b _nij > 0.

где S_n - число элементов изображения, которые составляют изображение объекта Н_n,
i - координата элемента изображения 1_nij по оси х, классифицированного как объект (b_nij>0),
j - координата элемента изображения l_nij по оси у, классифицированного как объект (b_nij>0).The coordinates λ _nx , λ _{ny of the} center of gravity of the binary image are calculated

where S _n is the number of image elements that make up the image of the object H _n ,
i is the coordinate of the image element 1 _nij along the x axis, classified as an object (b _nij > 0),
j is the coordinate of the image element l _nij along the y axis, classified as an object (b _nij > 0).

Limit the values of elements b _nij and write them to the address b _{(n-1) ij}
b _{(n-1) ij} = 32 if b _nij > 32,
b _{(n-1) ij} = b _nij if -32 <b _nij <32,
b _{(n-1) ij} = -32 if b _nij <-32.

If the number Sn of image elements classified as an object has exceeded the threshold value S ₀ , then a decision is made to detect the object in the frame, and the reading of the detection counter is increased by 1.

и скорости

объекта определяют в соответствии с выражением
по Х:

где

прогнозируемое на момент получения n-го кадра положение объекта (априорная оценка), рассчитываемое в соответствии с выражением (9);

апостериорная оценка положения объекта;

априорная оценка скорости объекта;

апостериорная оценка скорости объекта;
по Y:

где

прогнозируемое на момент получения n-го кадра положение объекта, рассчитываемое в соответствии с выражением (13);

апостериорная оценка положения объекта;

апостериорная оценка скорости объекта;

Если в текущем кадре объект не обнаружен, то координаты и скорость объекта прогнозируют в соответствии с выражениями

На четвертом этапе сглаживают изображение фона, для чего для всех точек изображения, классифицированных как фон (b_nij < 0), выполняют

где

0<k3<1.At the third stage, if they decide to detect an object in the frame (S> S ₀ ), then the coordinate estimates

and speed

the object is determined in accordance with the expression
on X:

Where

the position of the object predicted at the time of receiving the nth frame (a priori estimate) calculated in accordance with expression (9);

a posteriori assessment of the position of the object;

a priori estimate of the speed of an object;

a posteriori estimate of the speed of the object;
on Y:

Where

the position of the object predicted at the time of receiving the nth frame, calculated in accordance with expression (13);

a posteriori assessment of the position of the object;

a posteriori estimate of the speed of the object;

If the object is not detected in the current frame, then the coordinates and speed of the object are predicted in accordance with the expressions

At the fourth stage, the background image is smoothed, for which, for all image points classified as background (b _nij <0),

Where

0 <k3 <1.

где

0<k4<1.At the fifth stage, the reference image of the object is smoothed, for which, for all image points classified as an object (b _nij > 0),

Where

0 <k4 <1.

для обнуления элементов памяти, которые в предыдущих кадрах соответствовали объекту.In the case when the image point in the previous frame belonged to the object, and in the current frame is classified as background, smoothing the reference image of the object is performed in accordance with the expression

to reset the memory elements that in the previous frames corresponded to the object.

 At the sixth stage, if the value of the detection counter becomes 16, then they make a decision on tying the path and continue to perform calculations, otherwise they proceed to the steps in step 2, formula (5) and further.

изображения L_(n+1), центр которого совпадает с центром объекта:

(26)
где

элементы прогнозируемого участка изображения L_(n+1) размером МхМ,

сглаженное изображение объекта в соответствии с (24),

сглаженное изображение фона в соответствии с (23),
z_(n+1),ν,μ - бинарная маска объекта, сформированная по правилу
z_(n+1),ν,μ = 1, если b_(n+1)i1j1 > 0,
z_(n+1),ν,μ = 0, если b_(n+1)i1j1 ≤ 0,

где

определяются из (17) и (21).Form a predicted area

image L _{(n + 1)} , the center of which coincides with the center of the object:

(26)
Where

elements of the predicted image section L _{(n + 1)} of the MxM size,

smoothed image of the object in accordance with (24),

smoothed background image in accordance with (23),
z _{(n + 1), ν, μ} is the binary mask of the object, formed by the rule
z _{(n + 1), ν, μ} = 1 if b _{(n + 1) i1j1} > 0,
z _{(n + 1), ν, μ} = 0 if b _{(n + 1) i1j1} ≤ 0,

Where

are determined from (17) and (21).

сформированного в предыдущем кадре, по параметру λ_nx:

где

определяют по (26),

На восьмом этапе производят нелинейное оценивание координаты и скорости объекта
по Х:

где

прогнозируемое на момент получения n-го кадра положение объекта (априорная оценка), рассчитываемое в соответствии с выражением (9),

апостериорная оценка положения объекта,

априорная оценка скорости объекта,

апостериорная оценка скорости объекта.In the seventh step, the derivative of the predicted image is calculated

formed in the previous frame, by parameter λ _nx :

Where

determined by (26)

At the eighth stage, nonlinear estimation of the coordinate and speed of the object is performed
on X:

Where

the position of the object predicted at the time of receiving the nth frame (a priori estimate) calculated in accordance with expression (9),

a posteriori assessment of the position of the object,

a priori estimate of the speed of an object,

a posteriori estimate of the speed of the object.

сформированного в предыдущем кадре, по параметру λ_ny:

где

определяется (26),

На десятом этапе производят нелинейное оценивание координаты и скорости объекта по Y:

где

прогнозируемое на момент получения n-го кадра положение объекта (априорная оценка), рассчитываемое в соответствии с выражением (13),

апостериорная оценка положения объекта по Y,

априорная оценка скорости объекта по Y,

апостериорная оценка скорости объекта по Y.In a ninth step, a derivative of the predicted image is calculated

formed in the previous frame, by parameter λ _ny :

Where

defined by (26),

At the tenth stage, nonlinear estimation of the coordinate and speed of the object by Y is carried out:

Where

the position of the object predicted at the time of receiving the nth frame (a priori estimate) calculated in accordance with expression (13),

a posteriori assessment of the position of the object in Y,

a priori estimate of the speed of the object in Y,

a posteriori estimate of the object’s speed in Y.

где l_nij - элементы текущего обрабатываемого изображения L_n,

элементы прогнозируемого изображения фона,

сглаженное эталонное изображение объекта,
D = 1, центр строба ψ_n/ имеет координаты

На двенадцатом этапе находят координаты центра тяжести бинарного изображения, полученного в результате (27), и заменяют ими найденные на восьмом и десятом этапах оценки

По откорректированным значениям

формируют прогнозируемые на (n+1)-й кадр априорные оценки положения и скорости объекта

Сглаживают изображение фона для всех точек строба ψ_n размером МхМ текущего изображения L_n, классифицированных как фон (b_nij < 0):

где

0<k3<1.At the eleventh stage, the points of the portion (gate) ψ _n of the MxM size of the current image L _{n are} classified by checking the conditions

where l _nij are the elements of the current processed image L _n ,

predicted background image elements,

smoothed reference image of the object,
D = 1, the center of the gate ψ _{n /} has coordinates

At the twelfth stage, the coordinates of the center of gravity of the binary image obtained as a result of (27) are found and they are replaced by the estimates found at the eighth and tenth stages

Adjusted Values

form a priori estimates of the position and speed of the object predicted on the (n + 1) th frame

Smooth the background image for all points of the gate ψ _{n with the} size MxM of the current image L _n , classified as background (b _nij <0):

Where

0 <k3 <1.

где

0<k4<l.Smooth the reference image of the object for all points of the gate ψ _{n with the} size MxM of the current image L _n , classified as an object (b _nij > 0):

Where

0 <k4 <l.

где

,
для обнуления элементов памяти, которые в предыдущих кадрах соответствовали объекту.In the case when the image point in the previous frame belonged to the object, and in the current frame is classified as background, smoothing is performed in accordance with the expression

Where

,
to reset the memory elements that in the previous frames corresponded to the object.

 The value of Sn of the area of the object is controlled according to the classification results at the eleventh stage according to the formula (27).

 If Sn> S0, then perform the steps in the sixth stage and beyond.

 If Sn ≤ S0, then perform the steps in the second stage and further.

 A sign of the method that coincides with the sign of the proposed technical solution is the smoothing of the background image in accordance with expression (28).

 The reasons that impede the achievement of the required technical result when using the described method are as follows.

 The formation of the current coordinates of the object according to the results of processing the previous frame, essentially introduces an additional delay of up to 20 or 40 ms into the control loop of the position of the field of view depending on how the processing of television images is carried out - in half frames or frames, which limits the ability to determine coordinates and retention in the center of the field of view of dynamic objects.

 The classification of image elements into object and background points determined by the method at the second stage in accordance with rule (5) in the presence of noise leads to an erroneous determination of the coordinates of the object in the frame or analysis window, the dimensions of which significantly exceed the dimensions of the object (which is typical for small objects).

 The causes of these errors are as follows. The center of gravity of the binary image created by noise emissions does not depend on the coordinates of the captured object and is located near the center of the frame. The method does not provide for discarding points falsely classified in the second stage as an object due to noise emissions, therefore, if the object is not located in the center of the frame (for example, offset by 2-3 sizes), and the number of falsely classified image points is equal to or greater than the number points of the object, then the estimate of the coordinates of the object obtained by this method does not lie within the contour of the object, which leads to erroneous prediction of the position of the window for analyzing the image of the object in the next frame and the impossibility of further determination I had the coordinates.

 The method does not take into account the movement of the field of view during the time between obtaining images of adjacent frames (half-frames), which does not allow one to form a standard background, an average image of an object, and a predicted image of an object and background when the system is operating in a moving field of view.

 This is because for the formation of the reference background and the reference object, images of not one frame, but a sequence of frames are used, and during the formation of the mentioned standards, images of different frames must be shifted (and, in a more general case, rotated) relative to each other so that they are aligned the coordinates are the same background points (for the background standard) or object (for the object standard).

 To determine the necessary displacements (and, if necessary, rotations) of the images, all movements of the field of view must be taken into account.

The method uses the detection of all statistically significant deviations of the brightness of the elements of the current image from the a priori expected brightness of the background and object elements, regardless of the reasons for their occurrence. This leads to the inoperability of the method in the following conditions:
- during the movement of the carrier of the observation system and determining the coordinates of the object due to the selection of moving elements of the background in the image (if there is a moving object in the analysis field, this leads to erroneous coordinates of the object, and if there is a stationary object in the background, it is not detected);
- during the movement of the system carrier due to the presence of rotational movements of the field of view, since incorrect formation of the reference averaged background occurs, as mentioned above;
- in the presence of uncontrolled translational movements of the field of view (uncompensated stabilization errors or errors of the field of view drives) and a contrasting motley background.

 Under these conditions, the probability of a false classification of background elements as elements of an object sharply increases, which is the reason for the sharp increase in errors in determining the coordinates of an object and the impossibility of holding it in the center of the field of view.

 A known method of detecting moving objects and tracking them based on an analysis of the sequence of images (see patent 2 680 931 France, MKI 5 H 04 3/00, Procédé de detection et de suivi d'objets en mouvement par analyse de séquences d'images / Pineau Patrick et Robert Philippé (FR), THOMSON-CSF (FR), 91 10628; Declaration 08.28.91; Publ. 05.03.93, Bull. 09), consisting in the detection of moving objects, separation of related moving objects, prediction and time control position and movement of objects.

 The image is classified into three classes representing the background, the moving object and the shadow, by comparing the brightness of the image of the standard and the current image, the image of the standard is periodically updated and according to the method, the image points are classified into five classes representing the background contours, object contours in motion, background contours, hidden by a moving object, the background contours located in the shadow zone, and onto points that do not belong to any contour, and put a double label on each image point information, including information on the classification of points by comparing the brightness between the image of the standard and the current image, and contour information, taking into account the classification of points by comparing the contours of the image of the standard and the current image, and classifying image points into seven classes, each of which corresponds to one of seven types of labels with double information of the region (segment) and contour, for which the relaxation method is used with modeling the Markov field of labels with double information for optical ization and for classifying the pixels into seven classes of objects in the detection step.

 After the detection of moving objects, separation of related moving objects is performed, including relaxation with modeling of the Markov field of motion, temporary prediction of masks to optimize the classification of points into seven classes at the stage of detection of objects, and temporary prediction of motion to optimize the separation of related objects.

 This method allows you to detect moving objects and track them in a fixed field of view.

The signs of the method, coinciding with the signs of the proposed technical solution:
- detection of moving objects by comparing the brightness of the image of the background standard and the current image;
- periodic updating of the background reference image.

 The reasons that impede the achievement of the required technical result are the processing of the image to extract the image of the object from the background using the operations of detecting moving image segments, which leads to the inoperability of the method when observing an object that is stationary relative to the far background, and the presence of the carrier’s own movement of the tracking system, leading to the appearance of a moving near background, and with a contrasting background in the presence of uncontrolled movements of the field of view (uncompensated errors stabilization mode), since in this case moving segments of the background are allocated as an object, which does not allow detecting the true object.

 There is a method of recursively estimating displacements for tracking objects by analyzing sequences of digital images (see patent 0527791 EPO, MKI 5 G 06 F 15/70. VERFAHREN ZUR ZEITREKURSIVEN BEWEGUNGS-SCHATZUNG UND ZUR VERVOLGUNG BEWEGTER OBJEKELERERERITER DIEGEREJERGELERERER DIER BEER Klaus-Peter, Dr., Strassbergerstrasse, 8, W-8000 Munchen 40 (DE), SIEMENS AKTIENGESELLSCHAFT, Wittelsbacherplatz, 2, D-80333 Munchen (DE), EP 90108666; Claim 08.05.90; Published 20.07.94, Bull. 29; international priority PCT / EP91 / 00769; international publication WO 91/17513 11/14/91, Bull. 26).

c) определяют корректируемые переменные состояния (Y) из набора предсказанных переменных состояния (X) с помощью параметров измеренного перемещения (Z);
d) определяют предсказанные переменные состояния (X) для последующего момента из корректируемых переменных состояния (Y) с помощью рекурсивного фильтра, в котором:
k - индекс времени,
р - координаты пикселей,
j - сегментный индекс,
S(j,k,p) - сегментная функция перемещения сегмента j во время k в пикселе р,
s(n, k+l, T(p)) - величина сегментной функции перемещения сегмента n во время k+1 в пикселе Т(р),
l(k,p) - яркость изображения во время k в пикселе р,
l(k+l,T(p)) - яркость изображения во время k+1 в пикселе Т(р),
Т(р) - координаты пикселя р в результате преобразования Т.The method consists of the following set of actions:
- at startup, set the initial values of the predicted variables of the state of the object (X),
- for each image l (k) of a sequence of digital images in time l (1), l (2), ..., l (k):
a) detect the movement of segments in the image l (k) and these moving segments are determined by a binary segmenting function (S), which has the description
S (j, k, p) = 1 if the pixel p at time moment k belongs to the moving segment j,
S (j, k, p) = 0, otherwise;
b) measure the displacement parameters (Z) for these moving segments by maximizing the similarity function (C), which is generated using a set of displacement parameters (Z), transformations (T) from the segment functions (S) and digital images (I) in accordance with general formula

c) determine the corrected state variables (Y) from the set of predicted state variables (X) using the measured displacement parameters (Z);
d) determine the predicted state variables (X) for the next moment from the corrected state variables (Y) using a recursive filter in which:
k is the time index,
p is the coordinates of the pixels,
j is the segment index,
S (j, k, p) is the segment function of moving segment j during k in pixel p,
s (n, k + l, T (p)) is the value of the segment function of moving the segment n during k + 1 in the pixel T (p),
l (k, p) is the brightness of the image during k in pixel p,
l (k + l, T (p)) is the brightness of the image during k + 1 in the pixel T (p),
T (p) - the coordinates of the pixel p as a result of the transformation T.

 The method assumes that the family of transformations (T) includes either only affine transformations of a flat image, or only transfers (translational displacements).

 The state variables used (X, Y) describe the instantaneous positions, instantaneous velocities and accelerations of moving segments.

когда семейство преобразований (Т) включает аффинные преобразования плоского образа, или же в соответствии с формулой

когда семейство преобразований (Т) включает только переносы, где v - вектор смещения в плоскости образа.The similarity function (C) is formed in accordance with the formula

when the family of transformations (T) includes affine transformations of a flat image, or in accordance with the formula

when the family of transformations (T) includes only transfers, where v is the displacement vector in the image plane.

The signs of the method, coinciding with the signs of the proposed technical solution:
- detecting moving segments in the current image 1 (k) and determining these moving segments with a binary function (S) having the description
S (j, k, p) = 1 if the pixel p at time k belongs to the segment j,
S (j, k, p) = 0, otherwise.

 The reasons that impede the achievement of the desired technical result are the processing of the image to extract the image of the object from the background using the detection operations of moving image segments, which leads to the inoperability of the method when observing an object stationary relative to the far background, and the presence of the carrier’s own movement of the tracking system, leading to the appearance of a moving near background, as well as with a contrasting background in the presence of unpredictable movements of the field of view (uncompensated stabilization errors), since in this case moving segments of the background are allocated as an object, which does not allow to detect the object itself.

There is a method of tracking an object (see. Television tracking system with a Bayesian discriminator / Dr. T. Bakut, Ph.D. Labunets V.G. // "Foreign Radio Electronics." - 1987. - 10 . - Pages 81-93, - Rus.), Containing the following set of operations:
- input video image,
- analog-to-digital conversion of the video signal of the image and recording it in digital form in the buffer memory,
- a histogram classification of image points into object and background points based on a Bayesian decision rule (formation of a binary image of an object),
- destruction of the image elements of the object,
- determination of the vector shape of the areas of the object,
- determination of horizontal and vertical projections of the current binary image in the analysis window, as well as projections of the reference binary image of the object,
- definition of disjunctive convolution of the current and reference projections,
- determination of the coordinates (and speed) of the object by the position of the maximum of the obtained values of the disjunctive convolution of the current and reference projections,
- selection of the object from the image of the scene by discarding those image elements of the region of the object that are not included in the reference mask of the object, which is at the maximum position of the disjunctive convolution of the projections,
- prediction of the position of the object and the spatial strobe of the object in the next frame,
- re-detection of the object using the operations of the histogram classification of image points and operations to determine the center of the binary image of the object,
- development of control actions.

 When using this method for generating a binary image of an object, an adaptive, generally multi-threshold, statistically optimal rule for classifying image points into object and background points is implemented in the analysis window. The histogram classification of image elements in the analysis window is implemented by the following set of operations.

 In the current image frame, an analysis window is set, the center of which is located at the point of the predicted position of the object. Inside the analysis window, a spatial strobe of the object (PSO) and a background frame are formed, as shown in FIG.

Histograms G _PSO (L) and G _RF (L) of the brightness distribution L of the image in the spatial strobe of the object and inside the background frame, respectively, are calculated in the PSO windows and the background frame.

где b_ПСО, b_РФ - весовые коэффициенты, управляющие степенью усреднения гистограмм ПСО и рамки фона.To reduce fluctuations in point estimates of histograms G _PSO (L) and G _RF (L), the latter are averaged using first-order digital recursive filters:

where b _PSO , b _RF - weighting factors that control the degree of averaging of PSO histograms and background frames.

выполняют классификацию точек изображения в окне анализа в соответствии со следующим правилом.Using averaged histograms

carry out the classification of image points in the analysis window in accordance with the following rule.

где α - априорная вероятность наличия элементов фона в ПСО,
А - параметр стоимости неправильной классификации.An image point with coordinates (i, j) is considered to belong to the object, i.e., a binary image f (i, j) = l is formed if

where α is the a priori probability of the presence of background elements in the PSO,
A is the cost parameter of the incorrect classification.

 If condition (36) is not satisfied, then the point is considered to belong to the background and a binary image f (i, j) = 0 is formed.

 The points in the current frame are classified according to statistics from previous frames.

после чего определяют положения k₀ и 1₀ максимумов дизъюнктивных сверток S_г(k) и S_в(1), которые и принимают за координаты объекта в текущем окне анализа.Determine the coordinates of the center of the binary object image, by calculating the horizontal T _r (i) and the vertical _T (j) of the projection of the current binary image f (i, j), and the horizontal T _EG (i) and the vertical T _ev (j) of the projection of the reference binary images of the object f _ec (i, j) (mask of the object) formed on the previous frames, and disjunctive convolutions S _g (k), S _in (1) of the corresponding projections of the current and reference binary images are calculated in accordance with the expressions

after which the positions k ₀ and 1 _{0 of the} maxima of the disjunctive convolutions S _g (k) and S _in (1) are determined, which are taken as the coordinates of the object in the current analysis window.

The object is selected from the scene image by discarding those image elements of the object region that are not included in the reference mask of the object in position (k ₀ and 1 ₀ ).

 Predict the position of the image of the object in the next frame.

The signs of the method, coinciding with the signs of the proposed technical solution:
- receiving and storing signals of the current field of the television image,
- histogram classification of image points into object and background points based on a Bayesian decision rule (formation of a binary image of a histogram classifier),
- the formation of horizontal and vertical projections of the current binary image of the histogram classifier,
- re-detection of the image of the object when it is temporarily lost,
- development of signals that control the position of the field of view of the camera.

 The reasons that impede the achievement of the required technical result when using the method are as follows.

 To classify image elements in the method, the differences of the average histograms of the brightness distribution in the PSO and inside the background frame are used, which allows one to take into account information only about the relative frequency of occurrence of certain brightnesses in the area of the object and the surrounding background background and does not take into account their spatial distribution ("spatial pattern "). Because of this, the method is not applicable in those cases when, at the start of the tracking process and determining the coordinates of the histogram, the distribution of the features of the object and background (distribution of the brightness of their images) are close, which often happens when tracking a motley object against a finely dispersed motley background.

 To reduce fluctuations in the estimates of the histogram of the distribution of background brightness, it is averaged using a first-order digital recursive filter. When an object moves over an inhomogeneous background, this leads to the fact that while the object passes the boundary of the background inhomogeneity, the averaged histogram does not reflect the true distribution of background brightness around the object. In turn, this leads to an increase in the number of background image elements erroneously classified as an object and then to the “sliding” of the spatial strobe of the object onto the background section.

Closest to the claimed one is a method for determining the coordinates of an object based on the integrated use of the processing results of a sequence of television images (see THASSID Optoelectronic Guidance System / Kolmogorov G.S., Kostromitina E.V., Luchina I.I., Maltsev A.P. // "Foreign electronics". -1987. 10. -S. 57-68. -Eng.), Containing the following set of actions:
- input of the image video signal (receiving the current image field from the video camera),
- adaptive two-level quantization of the image video signal (generating a binary image of the object), followed by tracking the center of the binary image of the object inside the adaptive window (tracking system for the center of the object in the adaptive window - SSCAO),
- adaptive four-level quantization of the image video signal with the following:
1) the choice of landmarks,
2) by calculating the correlation functions by the method of absolute differences for each of the landmarks,
3) the determination of the position of the guidelines for the minimum correlation function,
4) by calculating the correlation function by the method of absolute differences for the object,
5) determining the position of the object by minimizing the correlation function (correlation tracking system - KSS),
6) rewriting of reference standards and an object,
7) image stabilization of a dynamic scene based on the use of the results of correlation tracking of landmarks,
8) selection of a moving object (formation of a binary image of a moving object from the image of the scene),
9) tracking the center of the binary image of a moving object (tracking system for a moving object - DSS),
10) electronic expansion of the field of view (scaling of the original image),
- assessment of tracking quality for each of the 3 groups of tracking operations (tracking the center of a binary image of an object inside an adaptive window, tracking the center of a binary image of a moving object, correlation tracking),
- selection of one of 3 groups of tracking operations (tracking the center of a binary image of an object inside an adaptive window, tracking the center of a binary image of a moving object, correlation tracking) to estimate the coordinates and speed of an object in accordance with established priorities:
1st priority (high) - tracking the center of the binary image of the object inside the adaptive window (group of operations of the Northwest Administrative District),
2nd priority - tracking the center of the binary image of a moving object (SDO operations group),
3rd priority (low) - correlation tracking of an object (KCC operations group) and setting the tracking operations group with the lowest priority of the object coordinates from the tracking operations group with the highest priority, if the coordinates of the objects formed by them differ significantly,
- determination of the coordinates and speed of the object,
- prediction of the position of the object and the spatial strobe of the object in the next frame,
- extrapolation of the coordinates of the object when the disappearance or rapid change of its image,
- re-detection of the object using the operations of adaptive two-level quantization of the video image signal and the operations of determining the center of the binary image of the object,
- development of control signals for the rotation of the field of view of the camera.

The known method of tracking an object allows you to:
- monitor the object with sufficient global contrast of the image of the object relative to the background of the terrain (light object on a dark background or a dark object on a light background), since the method ensures the correct classification of image points into points of the object and background points (see figure 3);
- monitor the object during its partial shielding by contrasting obstacles relative to the object (screening a light object against a dark background with dark objects or screening a dark object against a light background with light objects);
- monitor the object in the presence of a change in the image scale in the field of view over a wide range, as well as in the presence of relatively smooth (low-frequency) translational and rotational movements of the field of view;
- re-detect the contrast of the object during its short-term disappearance using the operations of adaptive two-level quantization of the video signal of the image and the operations of determining the center of the binary image of the object;
- monitor a fast-moving object in the absence or reduction of its global contrast relative to the background and the presence in the image of the object of local contrasting areas (spatial "pattern" of brightness).

The signs of the method, coinciding with the signs of the proposed technical solution:
- receiving and storing signals of the current field of the television image, the formation of the current image analysis window,
- selection on the image of fixed landmarks,
- determining the displacements of the landmarks of the current image frame relative to the position of the landmarks in the previous image frame by calculating the correlation functions of the landmark images,
- determination using the found displacements of the landmarks of the parameters of the shift and rotation of the images during the time between reception of the image fields,
- stabilization of the television image (transformation of the coordinates of the image elements with compensation for the shift and rotation of the image),
- scaling and storing the image analysis window of the current frame,
- subtraction from the image of the current frame of the image of the previously stored frame and the formation of the binary image of the selector of moving objects,
- remembering the reference image of the object,
- determination of the measure of dissimilarity of the current image and the reference image of the object in the form of a difference correlation function,
- use for the formation of control signals for the movement of the field of view of the video camera extrapolated coordinates of the object when the disappearance or rapid change of the image of the object.

 The reasons that impede the achievement of the required technical result are as follows.

 Firstly, operations of adaptive two-level quantization of an image video signal, providing single-threshold image binarization, correctly classify image points into object and background points only in a narrow range of conditions when there is a light object on a dark background or a dark object on a light background.

 As a result of this, under other conditions, in the presence of a significant number of erroneously classified image points (see Fig. 4), the PSO is gradually shifted to the background object. The adaptation operations of the quantization threshold of the video signal provide a change of this threshold to the side, which provides the best selection (binarization) of this background object, which leads to a further shift of the PSO to the background object, that is, to disrupt tracking. Moreover, the group of operations of adaptive two-level quantization of the video signal of the image and tracking the center of the binary image of the object inside the adaptive window described in the method does not have other means of distinguishing between image elements of the background and the object, except for the analysis of the binary image in the PSO. In a number of situations, this does not allow developing a sign of false tracking and, thus, transferring object tracking to other groups of operations within the framework of an integrated system. For this reason, the likelihood of a PSO “sliding” onto background elements increases significantly when moving an object over areas of the background that have brightnesses close to the brightness of the image of an object, especially when an object enters behind a screening background object due to a decrease in the area of the visible part of the object.

 Secondly, an erroneous estimation of the object’s speed due to the lag of the center of the visible part of the object from its true center when approaching the screening obstacle leads to the formation of incorrect predicted coordinates of the object and, as a result, erroneous signals to control the position of the camera’s field of view and erroneous positions and sizes of image analysis windows , which reduces the likelihood of re-detection of the object after its exit due to the background covering the object.

 Thirdly, when screening or a sharp change in the image of landmarks during the tracking process, for example, by smoke emissions, the operation of image stabilization operations using the results of correlation tracking of three landmarks is impossible, as a result of which the likelihood of maintaining tracking of the object is sharply reduced.

 Fourth, the method of stabilizing the image frame by the position of the landmarks is ineffective in the presence of broadband random movements, for example, a jitter in the field of view and the formation of a frame of a television (or thermal) image from several fields. This is due to the fact that the method implicitly assumes that during the formation of the image frame the spatial position of the field of view does not change. When the field of view flickers, this assumption is violated and, as a result, the relative position of the landmarks in different fields of the frame (or in different frames) used in the method for determining the parameters of affine transformations of an image differs from the true one.

 Fifthly, the position of the object described in the method of the correlation search operation does not provide a correct determination of the coordinates of the object when contrasting background elements fall into the image standard of the object, which is also one of the reasons for tracking failures.

 Sixth, the construction of a group of operations for selecting a moving object, which provides the formation of a binary image of a moving object, based on inter-frame subtraction of the current and previously memorized (reference) images, spatially stabilized using landmarks. The formation in the prototype method of a reference image directly from the current (with a delay of N frames, N = 1,2, ...) in the presence of noise or fluctuation of the background image of the terrain leads to a false classification of a large number of scene image elements as belonging to a moving object. This does not allow tracking a moving object of small sizes, the number of points in the image of which is commensurate with the number of falsely binarized elements of the terrain background.

 In addition, due to the third and fourth reasons, this group of operations is inoperative in situations where sufficiently accurate image stabilization using fixed landmarks is not provided.

 Seventhly, the selection at each moment of time during the data tracking interval of only one of the groups of tracking operations and the transfer of the coordinates of an object from a group with a higher priority to a group with a lower priority. The choice of one of the groups of tracking operations in this method is determined by the a priori assigned system of their priorities and the signs of reliability of tracking formed by these groups. However, the signs of the reliability of tracking each individually, as a rule, are not reliable enough in difficult observation conditions, as a result of which the incorrect coordinates of the object are imposed on those groups of tracking operations that are able to generate more accurate coordinates. An example of this is the situation of moving an object in the background of a locality having sections whose brightness is close to the brightness of the object. In this case, in the same way as it was described when considering the first reason, the PSO “crawls” against the background of the terrain, and due to the fact that this process is slow, and the density of the binary image due to the adaptation of the threshold remains high, a sign of reliability of tracking SSCAO signals the presence of confident tracking by this group of operations. Due to the high priority of the group of operations of adaptive two-level quantization of the video signal of the image and tracking the center of the resulting binary image of the object inside the adaptive window, when this method is implemented, the error coordinates of the object are imposed, for example, the correlation tracking system, which under these conditions can provide the determination of coordinates with the necessary accuracy.

 A device for determining the coordinates of an object (see the description of the invention to the copyright certificate 1562980 USSR, MKI 5 H 04 N 7/18, Device for determining the coordinates of an object. B.A. Alpatov, P.A. Bakut, I.E. Vornovitsky, A. A. Selyaev, A. I. Stepashkin, S. Yu. Khludov (USSR), - 4373608 / 24-09; Declaration 02.18.88; Publish. 07.05.90, Bull. 17), which contains a memory block of the signal of the smoothed image , a memory block of a signal of a reference image, memory blocks of a signal of a current image, an addressing block, a memory block of a weight function, a memory block of threshold functions, signal switching blocks, Lok calculating a difference signal of the current and reference images, a multiplier, a control unit, a coordinate calculating unit signals the object and the coordinate registers.

 This device allows under certain conditions to increase the accuracy of determining the coordinates of the object, because as the reference image, not the current image of the object is used, but a smoothed image of the object, on which the signal from the object is filtered from the additive noise distorting it. Using the weight function of the reference image can improve the accuracy of determining the coordinates of the object on a complex background, because when calculating the difference correlation function, the elements of the reference image that are not related to the desired object will not be used. In the device, as a reference, a smoothed image of the object is selected, the formation of which involves a large number of frames. In this case, the values of the sampling errors of the image of the object in the current and smoothed image can be considered independent, if during the smoothing period the object moves a distance of several discrete. Therefore, with such a periodic update of the standard, there will be no accumulation of an error in estimating the coordinates of the object. For the case of slow movement of an object in the device, the standard update algorithm is used, based on a comparison of the difference correlation function with a threshold, which also does not cause a rapid accumulation of errors in estimating the coordinates of the object. The device analyzes the movement of the object over a period of L frames, determined by the time it takes to establish the averaging filter standard, and depending on its size, one or another algorithm for updating the template is selected. This allows you to increase the accuracy of determining the coordinates of the object.

The reasons hindering the achievement of the required technical result when using this device are in determining the coordinates ν ^* , μ ^{* of the} image of the object based on minimizing the difference correlation function F (ν, μ) using the weight function Z (i, j in the first L + 1 frames ), equal to 1, which makes it impossible to correctly determine the coordinates of the object in conditions when the contrast of the object is lower than the contrast of the elements of the background surrounding it.

 The reasons for the inoperability of the device in the described situations are illustrated figa, 1b, 1c, 1d.

 In FIG. 1 shows the formation of a standard from the first received frame (figa) by cutting out its Central part (figb). At this stage, the standard is not yet protected from extraneous background elements. Typical for this situation, the distribution of the brightness of the image elements of the formed standard and the form of the difference correlation function F (ν, μ) are illustrated in Figs. 1c, 1d.

 As can be seen from FIG. With a contrasting motley background, the false minimum is smaller than the minimum corresponding to the position of the object. Thus, the estimation of the coordinates of an object using this device leads to an erroneous determination of coordinates and a breakdown in tracking.

 The reasons that impede the achievement of the required technical result can also include a combination of updating the standard and calculating the difference correlation function of the current and reference images, which leads to tracking failures when even relatively small areas of the object (10-20% of the area) overlap with high-contrast background elements.

 The reasons for the inoperability of the device in such situations are illustrated fig.1d.

 In the conditions under consideration, we can proceed from the fact that during the initial stage of tracking, the background along which the object was moving was such that the reference image of the object was formed correctly, that is, there are no background elements in the standard (Fig. 1d).

 When moving an object behind contrasting objects of the near background, as shown in Fig. 1f, the minimum value of the difference correlation function increases sharply, which causes a rewriting of the standard. As a result, the contrasting background element falls into the standard, and then, in the same way as described above, erroneous determination of the coordinates of the object and disruption of tracking occur.

 In addition, in the presence of unpredictable movements of the field of view during the rewriting of the pattern, background elements fall into the video window of the pattern of the object, which, as described above, also often leads to disruption of tracking.

 The signs of the device that match the features of the claimed technical solution: the combination of the unit for calculating the difference of the signals of the current and reference images, the multiplier by the weight function and the microprocessor of the unit for calculating the coordinates of the device object are equivalent to the former 60 (Fig. 11.1) of the block 11 of the claimed technical solution (Fig. 8.2) .

 Closest to the claimed is a device (see. Optical guidance system THASSID / Kolmogorov G.S., Kostromitina E.V., Luchina I.I., Maltsev A.P. // "Foreign radio electronics". -1987. 10. - P. 57-68. -Eng.), Which is a comprehensive tracking system, which includes SSCAO, SSSO, KSS, a microprocessor with the functions of a controller-dispatcher, combining all tracking systems (SSCAO, SSSO and KSS) into a single integrated tracking system. The input video data to the tracking systems comes from a camera installed on a gyro-stabilized platform, through adaptive quantizers (two-level - in SSCAO and four-level - in SSSO and KSS) and (if necessary) through an electronic expansion (scaling) block of the field of view.

 The microprocessor sets the mode and determines the operation of all tracking systems. He plays the role of the head controller. Its functions: the formation of control and intermediate data (coefficients, coordinates, thresholds, etc.) that determine the operation of the device in accordance with a given algorithm, for example, to control adaptive image quantization; to select landmarks by which correlation tracking is carried out; rewriting reference standards; image stabilization; control electronic expansion of the field of view, etc. According to the coordinates of the object coming from one of the monitoring systems in accordance with the established priority and depending on the observation area, the microprocessor controls the position of the gyrostabilized platform with a television camera tracking the monitoring object.

 Each tracking system works independently of the others until the coordinates of the object determined by it differ insignificantly from the coordinates determined by the tracking system with the highest priority. Otherwise, the lowest priority tracking system sets the coordinates of the object from the highest priority system. Thus, the systems operate in quasi-autonomous mode.

The main functions of the known device:
- adaptive quantization for better classification of scene image elements into object and background elements;
- implementation of the choice of scene segments as landmarks for correlation tracking;
- scene image stabilization due to correlation tracking of three landmarks;
- tracking the center of the binary image of the object inside the adaptive window;
- tracking a moving object using a subsystem that processes a stabilized image of the scene;
- correlation tracking of the object;
- correlation tracking of four fragments of the image of the object, in the final area of observation of the object.

 The reasons that impede the achievement of the required technical result are as follows.

 Firstly, adaptive two-level quantization of the video signal of the image, providing single-threshold binarization of the image, correctly classifies image points into points of the object and background only in a narrow range of conditions in the presence of a light object on a dark background or a dark object on a light background (see figure 3) .

 As a result of this, under other conditions, in the presence of a significant number of erroneously classified image points (see Fig. 4), the PSO is gradually shifted to the background object. Adaptation of the threshold of quantization of the video signal provides a change in this threshold to the side, which provides the best selection (binarization) of a given background object, which leads to a further shift of the PSO to the background object, that is, to disrupt tracking. Moreover, the means of adaptive two-level quantization of the video signal of the image and tracking the center of the binary image of the object inside the adaptive window used in the device do not have other possibilities for distinguishing between image elements of the background and the object, except for the analysis of the binary image in the PSO. In a number of situations, this does not allow developing a sign of false tracking and, thus, transferring object tracking to other groups of operations within the framework of an integrated system. For this reason, the likelihood of a PSO “sliding” onto background elements increases significantly when moving an object over areas of the background that have brightnesses close to the brightness of the image of an object, especially when an object enters behind a screening background object due to a decrease in the area of the visible part of the object.

 Secondly, an erroneous estimate of the speed of the object due to the lag of the center of the visible part of the object from its true center when approaching the screening obstacle (see Fig. 5) leads to the formation of incorrect predicted coordinates of the object and, as a result, erroneous signals to control the position of the camera’s field of view and erroneous position and size of the image analysis window, which reduces the likelihood of re-detection of the object after its exit due to the background covering the object.

 Thirdly, when screening or a sharp change in the image of landmarks during the tracking process, for example, by smoke emissions, stabilization of the scene image using the results of correlation tracking of three landmarks is impossible, as a result of which the likelihood of maintaining tracking of the object is sharply reduced.

 Fourth, the stabilization of the image frame by the position of the landmarks is ineffective in the presence of broadband random movements, for example, when the field of view is jittering and the formation of a television image frame from several fields. This is due to the fact that the device implicitly assumes that during the formation of the image field the spatial position of the field of view does not change. When the field of view is trembling, this assumption is violated and, as a result, the relative position of landmarks in different fields of the frame (or in different frames) used in the device to determine the parameters of affine image transformations differs from the true one.

 Fifth, the correlation search of the position of the object described in the device does not provide a correct determination of the coordinates of the object when contrasting background elements fall into the image region of the object standard (see Figs. 1c, 1d, 1e, 1e), which is also one of the reasons for tracking failures.

 Sixth, the implementation of selection operations of a moving object, providing the formation of a binary image of a moving object, based on inter-frame subtraction of the current and previously memorized (reference) images, spatially stabilized using landmarks. The formation of a reference image in the prototype device directly from the current one (delay by N frames, N = 1, 2, ...) in the presence of noise or fluctuations in the background image of the terrain leads to false binarization of a large number of scene image elements as belonging to a moving object. This does not allow tracking a moving object of small sizes, the number of points in the image of which is commensurate with the number of falsely binarized elements of the terrain background.

 In addition, due to the third and fourth reasons, the device’s DSS channel is inoperative in situations where sufficiently accurate image stabilization using landmarks is not provided.

 Seventhly, the device realizes the choice at each moment of time during the data tracking interval of only one of the groups of tracking operations and the transfer of the coordinates of the object from the group with a higher priority to a group with a lower priority. The choice of one of the groups of tracking operations in this device is determined by the a priori assigned system of their priorities and the signs of reliability of tracking formed by these groups. However, the signs of the reliability of tracking each individually, as a rule, are not reliable enough in difficult observation conditions, as a result of which the incorrect coordinates of the object are imposed on those groups of tracking operations that are able to generate more accurate coordinates. An example of this is the situation of moving an object in the background of a locality having sections whose brightness is close to the brightness of the object. In this case, in the same way as it was described when considering the first reason, the PSO “crawls” against the background of the terrain, and due to the fact that this process is slow, and the density of the binary image due to the adaptation of the threshold remains high, a sign of reliability of tracking SSCAO signals the presence of confident tracking by this group of operations. Due to the high priority of the group of operations of adaptive two-level quantization of the video signal of the image and tracking the center of the resulting binary image of the object inside the adaptive window, the implementation of this device involves the imposition of erroneous coordinates of the object, for example, a correlation tracking system, which under these conditions can provide the determination of coordinates with the necessary accuracy.

The features of the device, common with the features of the proposed technical solution, are as follows:
- the microprocessor of the device and the processor of the claimed technical solution - in terms of the formation of control and intermediate data (coefficients, coordinates, thresholds, etc.) that determine the operation of the device in accordance with a given algorithm;
- correlators of landmarks with buffer memory of the region of searching for landmarks and memory of standards of landmarks of the device and a block of stabilization parameters of the claimed technical solution - in terms of selecting landmarks on the image, determining the displacements of landmarks of the current image frame relative to the position of landmarks in the previous image frame by calculating the correlation functions of landmark images , definitions using the found displacements of the reference points of the parameters of the shift and rotation of the images during the time between reception her images.

 The objective of the invention is to expand the functionality of methods and devices for processing signals to determine the coordinates of objects observed in a sequence of television images, as well as to hold a television image of an object in the center of the field of view of the video camera of the surveillance system or in the center of the tracking window in a wide range of conditions for observing objects the presence of destabilizing factors without pre-tuning to the conditions of use with the participation of human operas torus.

 The technical result achieved by the implementation of the claimed invention is to ensure the determination of the coordinates and stable retention of the television image of the object in the center of the field of view of the video camera of the monitoring system or in the center of the tracking window when receiving from the video camera of the monitoring system both contrast and low-contrast television images of moving and stationary objects and the background of the terrain, when the object moves along the terrain, the average brightness of which coincides with the average brightness of the object , when changing the visible image of an object, due to its rotation during movement, changing illumination, observing the object under intermittent optical communications arising due to the obstruction of the object by various obstacles, such as folds of the terrain, buildings, vegetation, dust, smoke, splashes, etc. .p., with a controlled (mobile) and uncontrolled (fixed) field of view of the video camera of the surveillance system, in the presence of motion of the carrier of the monitoring system, causing uncontrolled movement of the field of view of the video camera system methods of observation (jitter, causing the effect of "blur" of the image, rotation of the field of view of the video camera of the observation system), as well as the appearance on the television image of a moving background of the area.

The technical result in terms of providing coordinates and stable image retention of the object in the center of the field of view of the video camera of the monitoring system or in the center of the tracking window when receiving from the video camera of the monitoring system both contrast and low-contrast images of moving and stationary objects and the background of the terrain, when moving the object across sections terrain, the average brightness of which coincides with the average brightness of the object, when the visible image of the object changes due to its rotation during movement, the change illumination, observation of an object under intermittent optical communications arising due to the obstruction of the object by various obstacles, such as folds of the terrain, buildings, vegetation, dust, smoke, splashes, etc., with controlled (moving) and uncontrolled (fixed ) the field of view of the video camera of the surveillance system, in the presence of movement of the carrier of the monitoring system, causing low-frequency uncontrolled movements of the field of view of the camera of the surveillance system (which lead to displacements and rotations of all elements of the field television image as a whole), as well as with the possible absence in the monitoring system of sensors of errors of stabilization of the field of view of the video camera, the monitoring system is achieved by the method of determining the coordinates of objects observed in a sequence of television images consisting in receiving and storing signals of the current field of the television image, the selection of signals from the current field of the television image of the signals of the television images 2N landmarks, and N = 3, 4, 5, ..., the formation of the signal s dissimilarity measure television pictures 2N guiding areas of the television image signals 2N guiding areas of the current field image and the television TV signals
images of the corresponding 2N landmarks of the terrain in the previous field of the television image and determining using them the offsets of the television images 2N landmarks of the terrain between the reception of signals from the current and previous fields of television images, determining using the offsets of the signals of the television images 2N landmarks of the terrain shift and rotation of the signals of the current field television images during the time between the reception of signals of the current and previous fields of television images, up to the signal path of the current n-th field of the television image, where n = 3, 4, 5, ..., determine the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of a television image, due to the action on the video camera of the monitoring system of the motion control signals of its field of view, determine the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system from the data of the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between the signals of the (n-1) -th and (n-2) -th fields of the television image, simultaneously with the reception of the signals of the current field of the television image, signals of uncontrolled movement and the roll of the field of view of the video camera of the monitoring system are received and used to generate signals of predicted coordinates television images of 2N landmarks in the current field of the television image, separate the signal-based dissimilarity measures of television images of 2N landmarks of the signal shift parameters in the current field of the television image during the time between the reception of signals of the current and previous fields of television images on the components of the controlled movement of the axis of the field of view of the video camera of the monitoring system and the uncontrolled movement of the axis of the field of view of the video camera of the monitoring system, the signals of the television image of the current n-th frame from the television image signals of the previous frame taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals (n-1) -go and (n-2) -th fields of the television image and signals of the current field of the television image with elementwise coordinate transformation of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the monitoring system, form the signals of the television image in the television image in the current analysis window and scale them using position and size signals of the current television image analysis window, gender resulting from processing the previous (n-1) -th frame of the television image, and the initial values of the position and size signals of the current analysis window of the television image are received from the monitoring system, the signals of the scaled television image are stored in the current analysis window, the signals of the differential television image of the moving selector are generated objects by subtracting from the signals of the scaled television image in the current window of the analysis of signals of the scaled television from images previously remembered and reduced to the current scale of the television image in the analysis window and shifted by the displacement of the axis of the field of view of the video camera of the surveillance system, form the signals of the primary binary television image of the selector of moving objects from the signals of the differential television image of the selector of moving objects, form the signals of the secondary binary television image selectors of moving objects from signals of a primary binary television image select a set of moving objects that have passed low-pass filtering, simultaneously with storing signals of a scaled television image in the current analysis window, generating signals of a differential television image of a selector of moving objects and generating signals of primary and secondary binary television images of a selector of moving objects from signals of a scaled television image in the current analysis window and taking into account the signals of the controlled movement of the axis of the field of view of the video camera system observation topics form the signals of the primary binary television image of the detector of background changes and the signals of the binary television image of the histogram classifier in the current analysis window, from the signals of the primary binary television image of the detector of background changes that have passed low-pass filtering, the signals of the secondary binary television image of the detector of background changes are formed, horizontal and vertical projections of signals of secondary binary television images the vector of moving objects and the detector of changes in the background and the horizontal and vertical projections of the signals of the binary television image of the histogram classifier, determine the reliability coefficients of the signals of the secondary binary television image of the detector of changes in the background and the selector of moving objects, as well as the signals of the binary television image of the histogram classifier, form a generalized horizontal and vertical projection signals of a generalized binary television image about an object from the horizontal and vertical projections of the signals of the secondary binary television images of the selector of moving objects and the detector of changes in the background, as well as from the horizontal and vertical projections of the signals of the binary television images of the histogram classifier based on their joint processing using the reliability coefficients of the signals of the binary television images of the histogram classifier, the signals of the secondary binary television images of the background change detector and the motion selector living objects, determine the horizontal and vertical boundaries, as well as the dimensions of the television image of the object according to the cutoff levels on the left and right, above and below the set percentage of the area of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, determine the current and average area of the generalized binary television image of the object located inside the formed boundaries of the television image of the object, determine the current coordinates in general the actual binary television image of the object, using the generalized horizontal and vertical projections of the generalized binary television image of the object, determine the current speed of the generalized binary television image of the object in the inertial coordinate system, determine the reliability coefficient of the current speed of the generalized binary television image of the object in the inertial coordinate system, simultaneously with the formation of signals of primary and secondary binary t the television images of the moving object selector and the background change detector, the binary television image of the histogram classifier, as well as the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object determine the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images as a result of nonlinear high-pass filtering of signals of scaled t the television image in the current analysis window, provided that the average area of the generalized binary television image of the object is exceeded by a threshold value, the received signals are stored, the signals of the static reference television image of the object or the signals of the static and dynamic reference television images of the object are brought, the signals of the static reference television image of the object are brought signals of static and dynamic reference television x the images of the object to the current scale, the formation and storage of signals of the measure of dissimilarity between the signals of the television images after non-linear high-frequency filtering of the signals of the scaled television image in the current analysis window and the signals of the static or static and dynamic reference television images of the object in the two-dimensional region for finding the displacements of the television image of the object, determining minimum values of signals of measure of dissimilarity of television images along the lines and along the columns of the two-dimensional region for searching for offsets of the television image of the object, forming sequences of the minimum values of the signals of the measure of dissimilarity of the television images along the rows and along the columns of the two-dimensional region for searching for offsets of the television image of the object, determining the corresponding coordinates of the television image of the object in the analysis window depending on the type of sequence of the minimum values of the signals of the measure dissimilarities of the television images for this coordinate, namely through analytical approximation of the sequence of minimum values of signals of the measure of dissimilarity of television images by a fourth degree polynomial and determining the coordinate of the television image of the object as the position of the minimum of the approximating polynomial, subject to assigning the sequence of minimum values of signals of the measure of dissimilarity of television images to the type of sequences with two boundaries of areas of rapid growth of values of signals of measure of dissimilarity of television images near the position of its minim ma, or by determining the displacement of the boundary of the region of rapid growth in the values of signals of the measure of dissimilarity of television images relative to the center of the analysis window and forming the coordinate of the object in the analysis window as a value proportional to the displacement of the border of the region of the fast growth of values of signals of the measure of dissimilarity of television images provided that the sequence of minimum signal values is assigned measures of dissimilarity of television images to the type of sequences with a flat position neighborhood the minimum and the presence of one region of rapid growth in the values of the signals of the measure of dissimilarity of television images, or by generating the coordinate of the object equal to the coordinate of the center of the analysis window, subject to assigning the sequence of the minimum values of signals of the measure of dissimilarity of television images to the type of sequences with a flat neighborhood of the position of the minimum values of the signals of the measure of dissimilarity of television images in the entire area of the search for displacements of the television image of the object, determine the current speed the movement of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object obtained on the basis of the formation of the signals of the measure of dissimilarity of the television images, determine the reliability coefficient of the current speed of the television image of the object in the inertial coordinate system obtained on the basis of the formation of signals of the measure of the dissimilarity of the television images, form a comprehensive assessment of the current speed of movement of the television image of the object in the inertial coordinate system from the data of the estimation of the current speed of the television image of the object obtained on the basis of the formation of signals of the measure of dissimilarity of the television images and the estimation of the current speed of the object by the generalized projections of the signals of the generalized binary television image of the object, taking into account the reliability coefficients of the speed of the television image object and a priori restrictions on the speed of maneuvering the object, average the signal Comprehensive estimates of the speed of moving the television image of the object in the inertial coordinate system and store them, determine the coordinates of the television image of the object in the field of view of the video camera of the monitoring system by integrating the difference of the complex estimate of the speed of moving the television image of the object in the inertial coordinate system and the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system in an inertial coordinate system, and the initial coordinates of the television image When the object is placed in the field of view of the video camera, the monitoring systems and the tracking start signal are received from the monitoring system, simultaneously with the formation of signals of primary and secondary binary television images of the moving object selector and background change detector, a binary television image of a histogram classifier, generalized horizontal and vertical projections of the generalized binary television signals images of the object form M windows for analysis of the background, and M = 4, 5, 6, ..., along the perimeter of the analysis window, and determine project the signals of the binary television image of the histogram classifier in M windows of analysis of the background, determine the area and coordinates of the boundaries of the binary television images of the histogram classifier in M windows of analysis of the background from the received projections of the signals of the binary television image of the histogram classifier in M windows of analysis of the background, generate control signals for moving the field axis view of a video camera of a surveillance system using the coordinates of a television image of an object in the video field of view amers of the surveillance system obtained as a result of processing a television image in the current analysis window or using extrapolated coordinates and the speed of moving the television image of the object, depending on the results of the analysis of the current and average area of the generalized binary television image of the object, the current and average speed of moving the television image of the object, area and of the coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows moreover, the extrapolated speed of moving the television image of the object is formed on the basis of the analysis of the stored values of the averaged complex estimate of the speed of moving the television image of the object, the position and size signals of the analysis window of the television image are generated for the next frame using the coordinates of the television image of the object in the field of view of the video camera of the observation system and dimensions television image of the object obtained by processing the current frame RA, and the initial values of the position signal and the size of the television image of the object and the signal start tracking are received from the monitoring system.

где X_УПР[i] , Y_УПР[i] - сигналы управления перемещением оси поля зрения видеокамеры системы наблюдения по горизонтали и вертикали, сформированные в результате обработки i-го поля телевизионного изображения,
hx[i] - импульсная характеристика горизонтального привода оси поля зрения видеокамеры системы наблюдения, представляющая собой отклик приращения угловых координат оси поля зрения видеокамеры системы наблюдения за время между приемом i-го и (i-1)-го полей телевизионного изображения на воздействие сигнала управления по горизонтали, постоянного на интервале получения первого поля телевизионного изображения и равного нулю в другие моменты времени,
hy[i] - импульсная характеристика вертикального привода оси поля зрения видеокамеры системы наблюдения, представляющая собой отклик приращения угловых координат оси поля зрения видеокамеры системы наблюдения за время между приемом i-го и (i-1)-го полей телевизионного изображения на воздействие сигнала управления по вертикали, постоянного на интервале получения первого поля телевизионного изображения и равного нулю в другие моменты времени,
К - длина импульсной характеристики, представляющая собой число полей телевизионного изображения, по истечении которых модуль импульсной характеристики не превышает заданного уровня.The technical result is also achieved by the fact that before receiving the signals of the current n-th field of the television image, where n = 3, 4, 5, ..., the controlled movement dx [n] and dy [n] of the axis of the field of view of the video camera of the surveillance system is determined the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image horizontally and vertically, respectively, by calculating the convolution of control signals X _UPR [i], Y _UPR [i] by moving the field of view of the camera of the surveillance system with impulse characteristics hx [i] and hy [i] of its drives

where X _UPR [i], Y _UPR [i] - control signals for the movement of the axis of the field of view of the video camera of the surveillance system horizontally and vertically, formed as a result of processing the i-th field of the television image,
hx [i] is the impulse response of the horizontal axis of the field of view of the video camera of the monitoring system, which is the response of the increment of the angular coordinates of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the ith and (i-1) th fields of the television image on the influence of the control signal horizontally constant in the interval of obtaining the first field of the television image and equal to zero at other points in time,
hy [i] is the impulse response of the vertical axis of the field of view of the video camera of the surveillance system, which is the response of the increment of the angular coordinates of the axis of the field of view of the video camera of the surveillance system between the reception of the i-th and (i-1) th fields of the television image on the influence of the control signal vertically, constant in the interval of obtaining the first field of the television image and equal to zero at other points in time,
K is the length of the impulse response, which is the number of fields of the television image, after which the impulse response module does not exceed a predetermined level.

jy - номер строки в кадре, jy=1,...,Mк;
Nк - число элементов изображения в строке;
Мк - число строк в кадре изображения;
npk - индекс текущего полукадра:
npk = 1 - в нечетных полукадрах,
npk = 0 - в четных полукадрах,

Технический результат достигается также тем, что сигналы L_{1 БИН ДИФ}(ix, jy) первичного бинарного телевизионного изображения детектора изменений фона формируют приведением сигналов L_n-1эф(ix, jy) телевизионного изображения эталонного фона, полученных в предыдущем n-1 кадре, к текущему масштабу, образованием сигналов L_{р ДИФ}(iх,jу) разностного телевизионного изображения детектора изменений фона вычитанием из сигналов L_{n OA}(ix,jy) масштабированного изображения в текущем окне анализа сигналов L_n-1эф(ix,jy) телевизионного изображения эталонного фона предыдущего кадра со сдвигом, учитывающим перемещение Vx, Vy центра окна анализа в инерциальной системе координат за последний кадр:
L_{p ДИФ}(ix,jy) = L_{n ОА}(ix,iy) - L_{n-1 эф}(ix+Vx,jy+Vy),
определением порога бинаризации ПОРОГ _ДИФ(ix,jy) детектора изменений фона как величины, пропорциональной локальному параметру рассеяния значений сигналов разностного телевизионного изображения L_{р ДИФ}(iх, jу) детектора изменений фона в окрестности точки с координатами ix,jy, присвоением значений первичному бинарному телевизионному изображению L_{1 БИН ДИФ}(ix,jy) детектора изменений фона
L_{1 БИН ДИФ}(ix,jy)=1,
если |L_{p ДИФ}(ix,jy)|≥ПОРОГ_ДИФ(ix,jy),
или
L_{1 БИН ДИФ}(ix,jy)=0,
если |L_{p ДИФ}(ix,jy)|<ПОРОГ_ДИФ(ix,jy),
где ix, jy - координаты сигналов масштабированного телевизионного изображения относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
причем сигналы L_{n эф}(ix,jy) телевизионного изображения эталонного фона формируют разделением сигналов масштабированного телевизионного изображения в текущем окне анализа на сигналы телевизионного изображения трех типов:
сигналы телевизионного изображения в окне объекта - ОО_ДИФ,
сигналы телевизионного изображения в окне фона - ОФ_ДИФ,
сигналы телевизионного изображения в окне - "Новый фон" - НФ,
где в качестве сигналов телевизионного изображения в окне объекта определяют сигналы телевизионного изображения в прямоугольнике, находящемся в центре текущего окна анализа и включающем в себя преимущественно элементы изображения объекта, в качестве сигналов телевизионного изображения окна "Новый фон" определяют элементы изображения на внешних границах текущего окна анализа, на которых за счет движения объекта и перемещения оси поля зрения видеокамеры системы наблюдения появляются новые элементы телевизионного изображения фона, в качестве сигналов телевизионного изображения окна фона определяют все оставшиеся элементы телевизионного изображения окна анализа, запоминанием в окне "Новый фон" сигналов L_{n OA}(ix,jy) из текущего окна анализа масштабированного телевизионного изображения текущего n-го кадра:
L_{n эф}(ix,jy) = L_{n OA}(ix,jy), ix,jy ∈ НФ,
где L_{n OA}(ix,jy) - значения сигналов яркости элемента масштабированного телевизионного изображения в окне анализа с координатами ix,jy, усреднением в окне фона сигналов масштабированного телевизионного изображения L_{n OA}(ix, jy) из текущего окна анализа с постоянной Wоф и с учетом сдвига окна анализа в инерциальной системе координат за последний кадр:
L_{n эф}(ix, jy) = (l-Wоф)*L_{n-1 эф}(iх+Vx, jy+Vy)+Wоф*L_{n OA}(ix, jy) ix, je∈OФ_ДИФ,
где Vx, Vy - перемещение центра окна анализа за последний кадр по горизонтали и вертикали, соответственно, в инерциальной системе координат, перезаписью в окне объекта сигналов телевизионного изображения эталонного фона предыдущего кадра со сдвигом, учитывающим перемещение центра окна анализа за последний кадр:
L_{n эф}(ix,jy) = L_{n-1 эф}(iх+Vх,iy+Vy)
ix,jy ∈ ОО_ДИФ.The technical result is also achieved by the fact that the formation of signals L _FRAME (ix, jу) of the image of the current frame during interlacing from the signals L _P-FRAME (i, p, npk) of the current image half-frame and the image L _1-KADR (ix, jy) of the previous the frame is carried out by predicting the signals L _KADR (ix, jy) of the image of the current frame, using the image offset L _-1KAAD (ix, iy) of the previous frame by the amount of controlled movement dx, dy of the axis of the field of view of the camera of the horizontal and vertical surveillance system, respectively, for time between receiving p image lukadrov
L _KADR (ix, jy) = L _-1KAADR (ix + dx, jy + dy)
and replacing the image points of the current frame with the image points of the current half frame with compensation for the current uncontrolled displacements r _x , r _y and roll φ of the field of view of the video camera of the surveillance system
L _FRAME [ix (i, p, npk), jy (i, p, npk)] = L _P-KADR (i, p, npk),
where i is the number of the element in the image line of the current half frame, i = 1, ..., Nк;
ix - element number in the image line of the current frame, iх = 1, ..., Nк;
p is the line number in the half frame,

jy - line number in the frame, jy = 1, ..., Mк;
Nк - the number of image elements in a row;
Mk - the number of lines in the image frame;
npk - index of the current half-frame:
npk = 1 - in odd half frames,
npk = 0 - in even half frames,

The technical result is also achieved by the fact that the signals L _{1 BIN DIF} (ix, jy) of the primary binary television image of the background change detector are formed by converting the signals L _n-1eff (ix, jy) of the television image of the reference background obtained in the previous n-1 frame to the current scale, the formation of signals L _{r DIF} (ix, jy) of the differential television image of the background change detector by subtracting the scaled image from the signals L _{n OA} (ix, jy) in the current signal analysis window L _n-1eff (ix, jy) of the television image of the reference background pre yduschego frame shift, which takes into account movement of Vx, Vy analysis window center in the inertial frame for the last frame:
L _{p DIF} (ix, jy) = L _{n ОА} (ix, iy) - L _{n-1 eff} (ix + Vx, jy + Vy),
determination of the binarization threshold _DIF threshold (ix, jy) of the background change detector as a value proportional to the local parameter of the scattering of the signals of the difference television image L _{r DIF} (ix, jy) of the background change detector in the vicinity of the point with coordinates ix, jy, assignment of values to the primary binary television image L _{1 BIN DIF} (ix, jy) of the background change detector
L _{1 BIN DIF} (ix, jy) = 1,
if | L _{p DIF} (ix, jy) | ≥POROG _DIF (ix, jy),
or
L _{1 BIN DIF} (ix, jy) = 0,
if | L _{p DIF} (ix, jy) | < _DIF Threshold (ix, jy),
where ix, jy are the coordinates of the signals of the scaled television image relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
moreover, the signals L _{n eff} (ix, jy) of the television image of the reference background are formed by dividing the signals of the scaled television image in the current analysis window into three types of television image signals:
signals of the television image in the object window - OO _DIF ,
signals of the television image in the background window - OF _DIF ,
signals of the television image in the window - "New background" - NF,
where the signals of the television image in the window of the object determine the signals of the television image in the rectangle located in the center of the current analysis window and which includes mainly the image elements of the object, as the signals of the television image of the "New background" window, image elements are determined at the external borders of the current analysis window on which new elements of the television background image appear due to the movement of the object and the displacement of the axis of the field of view of the video camera of the surveillance system, achestve television image signals background windows define all remaining elements of the television image analysis windows, latching in the "New background" signals L _{n OA} (ix, jy) of the current window of analysis scaled television image of the current n-th frame:
L _{n eff} (ix, jy) = L _{n OA} (ix, jy), ix, jy ∈ NF,
where L _{n OA} (ix, jy) are the values of the brightness signals of the scaled television image element in the analysis window with coordinates ix, jy, averaging in the background window of the signals of the scaled television image L _{n OA} (ix, jy) from the current analysis window with the constant Wof and taking into account the shift of the analysis window in the inertial coordinate system for the last frame:
L _{n eff} (ix, jy) = (l-Woff) * L _{n-1 eff} (ix + Vx, jy + Vy) + Woff * L _{n OA} (ix, jy) ix, je∈OF _DIF ,
where Vx, Vy is the displacement of the center of the analysis window for the last frame horizontally and vertically, respectively, in the inertial coordinate system, overwriting in the window of the signal object of the television image the reference background of the previous frame with a shift that takes into account the displacement of the center of the analysis window for the last frame:
L _{n eff} (ix, jy) = L _{n-1 eff} (ix + Vx, iy + Vy)
ix, jy ∈ OO _DIF .

где ix, jy - координаты сигналов отфильтрованного телевизионного изображения S_fil_ДИФ(iх,jу) относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа;
NF, MF - параметры апертуры низкочастотного фильтра по горизонтали и вертикали, соответственно;
di, dj - внутренние переменные низкочастотного фильтра по горизонтали и вертикали, соответственно, di∈[-NF, NF], dj∈[-MF, MF];
h_HЧ[di, dj] - импульсная характеристика низкочастотного фильтра.The technical result is also achieved by the fact that the low-pass filtering of the signals of the primary binary television image L _{1 BIN DIF} (ix, jy) of the background change detector is carried out using two-dimensional convolution

where ix, jy are the coordinates of the signals of the filtered television image S_fil _DIF (ikh, jy) relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
di, dj are the internal variables of the low-pass filter horizontally and vertically, respectively, di∈ [-NF, NF], dj∈ [-MF, MF];
h _HF [di, dj] - impulse response of a low-pass filter.

где ix, jy - координаты сигналов отфильтрованного телевизионного изображения S_fil_СДО(ix,jy) относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа;
NF, MF - параметры апертуры низкочастотного фильтра по горизонтали и вертикали, соответственно;
di, dj - внутренние переменные низкочастотного фильтра по горизонтали и вертикали, соответственно, di ∈[-NF, NF], dj ∈[-MF, MF];
h_HЧ[di, dj] - импульсная характеристика низкочастотного фильтра.The technical result is also achieved by the fact that the low-pass filtering of the signals of the primary binary television image L _{1 BIN SDO} (ix, jy) of the selector of moving objects is carried out using two-dimensional convolution

where ix, jy are the coordinates of the filtered television image S_fil _DLS (ix, jy) relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
di, dj are the internal variables of the low-pass filter horizontally and vertically, respectively, di ∈ [-NF, NF], dj ∈ [-MF, MF];
h _HF [di, dj] - impulse response of a low-pass filter.

NX, NY - размеры текущего окна анализа;
Роrog_ДИФ1, Роrog_ДИФ0 - значения порогов принятия решения для единичных и нулевых элементов первичного бинарного телевизионного изображения детектора изменений фона соответственно.The technical result is also achieved by the fact that the signals L _{2BIN DIF} (ix, jy) of the secondary binary television image of the background change detector are formed from the signals of the low-pass filter S_fil _DIF (ix, jу) in accordance with the rule
L _{2BIN DIF} (ix, jy) = 1,
if S_fil _DIF (ix, iy)> Porog _DIF 1 and L _{1 BIN DIF} (ix, jy) = 1
or
S_fil _DIF (ix, jy)> Porog _DIF 0 and L _{1 BIN DIF} (ix, jy) = 0,
L _{2BIN DIF} (ix, jx) = 0 - otherwise,
where ix, jy are the coordinates of the signals L _{2BIN DIF} (ix, jy) of the secondary binary television image of the background change detector relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
Rorog _DIF 1, Rorog _DIF 0 - values of decision thresholds for single and zero elements of the primary binary television image of the background change detector, respectively.

NX, NY - размеры текущего окна анализа,
Porog_СДО1, Porog_СДО0 - значения порогов принятия решения для единичных и нулевых элементов первичного бинарного телевизионного изображения селектора движущихся объектов соответственно.The technical result is also achieved by the fact that the signals of the secondary binary television image L _{2 BIN SDO} (ix, jy) of the selector of moving objects are formed from the signals S_fil _SDO (ix, jy) of the low-pass filter in accordance with the rule
L _{2BIN LMS} (ix, jy) = 1,
if S_fil _SDO (ix, iy)> Porog _SDO 1 and L _{1 BIN SDO} (ix, jy) = 1
or S_fil _LMS (ix, jy)> Porog _DLS 0 and L _{1 BIN LMS} (ix, jy) = 0,
L _{2BIN LMS} (ix, jx) = 0 - otherwise,
where ix, jy are the coordinates of the signals L _{2 BIN SDO} (ix, jy) of the secondary binary television image of the selector of moving objects relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
Porog _SDO 1, Porog _SDO 0 - values of decision thresholds for single and zero elements of the primary binary television image of the selector of moving objects, respectively.

L _{БИН ГК}(iх,jу) = 0 - в противном случае,
где ix, jy - координаты сигналов бинарного телевизионного изображения L_{БИН ГК}(ix,jy) относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
L_{n OA}(ix, jy) - сигналы масштабированного телевизионного изображения в текущем окне анализа;
W_n-1 ^H _ЦОО[L_i] - нормированная гистограмма распределения яркостей L_i сигналов масштабированного телевизионного изображения в центральном окне объекта - ЦОО, полученная в предыдущем (n-1)-м кадре;
W_n-1 ^H _ОФГК[L_i] - нормированная гистограмма распределения яркостей сигналов телевизионного изображения фона в окне фона гистограммного классификатора - ОФГК, полученная в предыдущем (n-1)-м кадре;
L_i - уровень яркости сигналов телевизионного изображения;
i - номер уровня яркости сигналов телевизионного изображения, i = 1,..., N_ур;
α_n-1 - параметр, зависящий от числа элементов телевизионного изображения в окне объекта гистограммного классификатора - ООГК, классифицированных как фон в предыдущем (n-1)-м кадре;
A(ix,jy) - штрафная функция, зависящая от координат элемента ix,jy телевизионного изображения в окне анализа, в том числе при ix,jy ∈ ЦОО A(ix,jy) = А₀;
причем после определения координат Х_O, Y_O и размеров Rxo, Ryo телевизионного изображения объекта в текущем n-м кадре в качестве центрального окна объекта - ЦОО определяют прямоугольник с размерами Кх_ЦОО=Кхо, Ку_ЦОО=уо, находящийся внутри окна анализа, центр которого совпадает с центром Хо, Yo телевизионного изображения объекта и включающий в себя преимущественно элементы телевизионного изображения объекта, в качестве окна объекта гистограммного классификатора - ООГК определяют область, заключенную между двумя прямоугольниками с общим центром и с размерами Rх_ЦОО, Rу_ЦОО и Rx_ООГК, Ry_ООГК, где Rх_ООГК>Rх_ЦОО и Rу_ООГК > Ry_ЦОО, в качестве окна фона гистограммного классификатора - ОФГК определяют область, заключенную между двумя прямоугольниками с общим центром и с размерами Rx_ООГК, Ry_ООГК и Rх_ОФГК, Rу_ОФГК, где Rх_ОФКГ > Rх_ООКГ и Rу_ОФКГ> Ry_ООКГ, по сигналам телевизионного изображения L_{n OA}(ix, jy), считываемым из окна ОФГК в текущем n-м кадре, определяют гистограмму W_ОФКГ[L_i] распределения яркостей L_i изображения фона, по сигналам телевизионного изображения, выбираемым из окна ЦОО в текущем n-м кадре, определяют гистограмму W_ЦОО[L_i] распределения яркостей L_i изображения объекта, производят сглаживание гистограмм W_ОФКГ[L_i] и W_ЦОО[L_i]

где h_cг[j] - импульсная характеристика сглаживающего фильтра;
(2*ns+1) - число точек импульсной характеристики сглаживающего фильтра;
определяют текущий порог П_г усреднения сглаженной гистограммы

распределения яркостей изображения объекта в соответствии с выражением

где П_г0 и k_пг - постоянные величины;
1[х] - единичная функция, определяемая условиями:
1[х] = 1 при х≥0,
1[х]=0 при х<0;
усредняют текущий порог П_г рекурсивным фильтром первого порядка

где γ_{n пор} - постоянная фильтра усреднения порога П_г, изменяющаяся в зависимости от номера кадра от значения γ_{1 пор} = 1 в первом кадре до стационарного значения

n - номер текущего кадра изображения;
ограничивают значения усредненного порога сверху и снизу, усредняют сглаженную гистограмму

распределения яркостей L_i телевизионного изображения объекта рекурсивным фильтром первого порядка

где γ_{n ЦОО} - постоянная фильтра усреднения гистограммы распределения яркостей телевизионного изображения объекта, изменяющаяся в зависимости от номера кадра от значения γ_{1 ЦОО} = 1 в первом кадре до стационарного значения γ_ЦОО, причем, начиная с номера кадра, превосходящего Nку,
где Nку - номер кадра, с которого усреднение гистограммы

выполняют по условиям, Nку=16,...,128,
усреднение гистограммы

распределения яркостей L_i изображения объекта выполняют только для тех уровней яркости L_i, для которых одновременно выполняются два условия:

формируют нормированную W_n-1 ^H _ЦОО[L_i] гистограмму распределения яркостей изображения объекта из усредненной гистограммы

распределения яркостей L_i изображения объекта

где N_yp - число уровней яркости изображения L_{n OA}(ix,jy),
формируют нормированную W_n-1 ^H _ОФГК[L_i] гистограмму распределения яркостей изображения фона из сглаженной гистограммы

распределения яркостей Li изображения фона:

Технический результат достигается также тем, что горизонтальную и вертикальную проекции G_{пp СДО}(ix), V_{пр СДО}(jy) сигналов вторичного бинарного телевизионного изображения L_{2БИН СДО}(iх, jу) селектора движущихся объектов определяют в соответствии с

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где ix, jy - координаты сигналов L_{2БИН СДО}(iх,jу) вторичного бинарного телевизионного изображения селектора движущихся объектов относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа;
Nwin и Mwin - размеры вторичного бинарного изображения селектора движущихся объектов по горизонтали и вертикали соответственно.The technical result is also achieved by the fact that the signals L _{BIN GK} (ix, jy) of the binary television image of the histogram classifier are formed in accordance with the rule
L _{BIN GK} (ix, jy) = 1 if

L _{BIN GK} (ix, jy) = 0 - otherwise,
where ix, jy are the coordinates of the signals of the binary television image L _{BIN GK} (ix, jy) relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
L _{n OA} (ix, jy) - signals of the scaled television image in the current analysis window;
W _n-1 ^H _COO [L _i ] - normalized histogram of the distribution of brightness L _{i of the} signals of the scaled television image in the central window of the object - COO obtained in the previous (n-1) -th frame;
W _n-1 ^H _OFGK [L _i ] - normalized histogram of the distribution of the brightness of the signals of the television background image in the background window of the histogram classifier - OFGK obtained in the previous (n-1) -th frame;
L _i - the brightness level of the television image signals;
i is the number of the brightness level of the television image signals, i = 1, ..., N _ur ;
α _n-1 is a parameter that depends on the number of elements of the television image in the window of the object of the histogram classifier - OOGK, classified as background in the previous (n-1) -th frame;
A (ix, jy) is a penalty function that depends on the coordinates of the element ix, jy of the television image in the analysis window, including for ix, jy ∈ CCA A (ix, jy) = A ₀ ;
moreover, after determining the coordinates X _O , Y _O and the dimensions Rxo, Ryo of the television image of the object in the current nth frame, the rectangle with the dimensions Kx _COO = Kho, Ku _COO = yo inside the analysis window is determined as the center window of the object — COO, center which coincides with the center Ho, Yo of the television image of the object and including mainly the elements of the television image of the object, as the window of the object of the histogram classifier - OOGK define the area enclosed between two rectangles with a common center om and with the dimensions Rx _SSC, Ru _CSB and Rx _OOGK, Ry _OOGK where Rx _OOGK> Rx _CSB and Ru _OOGK> Ry _CSB, as a window background histogram classifier - OFGK define the area contained between the two rectangles with a common center and with dimensions Rx _OOGK , Ry _OOGK and Rx _OFGK , Rу _OFGK , where Rx _OFKG > Rx _OGKG and Rу _OFKG > Ry _OOGK , from the television image signals L _{n OA} (ix, jy) read from the OFGK window in the current nth frame, determine histogram W _OFCG [L _i ] of the brightness distribution L _{i of the} background image, according to the television image signals selected from the CCO window in the current nth frame, determine the histogram W of the _COO [L _i ] of the brightness distribution L _{i of the} image of the object, smooth out the histograms of W _OFCG [L _i ] and W of the _COO [L _i ]

where h _cg [j] is the impulse response of the smoothing filter;
(2 * ns + 1) is the number of points of the impulse response of the smoothing filter;
determine the current threshold P _g averaging the smoothed histogram

the brightness distribution of the image of the object in accordance with the expression

where P _g0 and k _pg are constant values;
1 [x] is the unit function determined by the conditions:
1 [x] = 1 for x≥0,
1 [x] = 0 for x <0;
average the current threshold P _g with a first-order recursive filter

where γ _{n pore} is the threshold averaging filter threshold P _g , which varies depending on the frame number from the value γ _{1 pore} = 1 in the first frame to a stationary value

n is the number of the current image frame;
limit the values of the averaged threshold above and below, average the smoothed histogram

the brightness distribution L _{i of the} television image of the object with a first order recursive filter

where γ _{n COO} is the filter averaging constant of the histogram of the brightness distribution of the television image of the object, which varies depending on the frame number from the value of γ _{1 COO} = 1 in the first frame to the stationary value of the _COO , and, starting with the frame number exceeding Nku,
where Nk is the frame number from which the histogram averaging

fulfill the conditions, Nku = 16, ..., 128,
histogram averaging

the brightness distribution L _{i the} image of the object is performed only for those brightness levels L _i for which two conditions are simultaneously satisfied:

form a normalized W _n-1 ^H _COO [L _i ] histogram of the distribution of the brightness of the image of the object from the average histogram

brightness distribution L _i image of the object

where N _yp is the number of image brightness levels L _{n OA} (ix, jy),
form a normalized W _n-1 ^H _OFGK [L _i ] histogram of the distribution of the brightness of the background image from the smoothed histogram

luminance distribution Li background image:

The technical result is also achieved in that the horizontal and vertical projections G _{prosp LMS} (ix), V _{ave LMS} (jy) of the secondary binary signal television image L _{2BIN LMS} (ix, ju) selector moving objects is determined according to

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where ix, jy are the coordinates of the signals L _{2 BIN SDO} (ix, jy) of the secondary binary television image of the moving objects selector relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
Nwin and Mwin are the sizes of the secondary binary image of the moving object selector horizontally and vertically, respectively.

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где ix, jy - координаты сигналов L_{2БИН ДИФ}(ix,jy) вторичного бинарного телевизионного изображения детектора изменений фона относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
Nwin и Mwin - размеры вторичного бинарного телевизионного изображения детектора изменений фона по горизонтали и вертикали соответственно.The technical result is also achieved in that the horizontal and vertical _straight projection _DIF G (ix), V _{ave DIF} (ju) of the secondary binary signal television image L _{2BIN DIF} (ix, ju) detector background changes determined in accordance with

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where ix, jy are the coordinates of the signals L _{2BIN DIF} (ix, jy) of the secondary binary television image of the background change detector relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
Nwin and Mwin are the sizes of the secondary binary television image of the background and horizontal changes detector, respectively.

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где ix,jy - координаты сигналов L_{БИН ГК}((ix,jy) бинарного телевизионного изображения гистограммного классификатора относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа;
Nwin и Mwin - размеры бинарного изображения гистограммного классификатора по горизонтали и вертикали соответственно.The technical result is also achieved by the fact that the horizontal and vertical projections G _{pr GK} (ix), V _{pr GK} (jy) of the binary image signals L _{BIN GK} (ix, jy) of the histogram classifier is determined in accordance with

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where ix, jy are the coordinates of the signals L of the _{BIN GC} ((ix, jy) of the binary television image of the histogram classifier relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
Nwin and Mwin are the sizes of the binary image of the histogram classifier horizontally and vertically, respectively.

причем усредненные плотности

бинарных телевизионных изображений детектора изменений фона, селектора движущихся объектов и гистограммного классификатора получают в результате ограничения минимальных и максимальных значений и последующего усреднения рекурсивными фильтрами первого порядка текущих плотностей V_ДИФ(n), V_СДО(n), V_ГК(n) соответствующих бинарных телевизионных изображений,
где

S_ОДИФ(n), S_ОСДО(n), S_ОГК(n) - текущие площади бинарных телевизионных изображений детектора изменений фона, селектора движущихся объектов и гистограммного классификатора, соответственно, внутри границ телевизионного изображения объекта;
S_OO(n) - текущая площадь области внутри границ телевизионного изображения объекта;
F_{ну_ДИФ}(n), F_{ну_СДО}(n), F _{ну_ГК}(n) - функции ввода начальных условий детектора изменений фона, селектора движущихся объектов и гистограммного классификатора;
n - номер текущего кадра.The technical result is also achieved by the fact that the reliability coefficients W _DIF , W _SDO , W _{GK are} defined as the product of the input functions of the initial conditions and the normalized average densities of binary television images of the background change detector, moving object selector and histogram classifier, respectively

moreover, the averaged densities

binary television images of the background change detector, moving object selector, and histogram classifier are obtained by limiting the minimum and maximum values and subsequent averaging of the first densities V _DIF (n), V _SDO (n), V _GK (n) of the corresponding binary television images
Where

S _ODIF (n), S _OSDO (n), S _OGK (n) - current areas of binary television images of the background change detector, moving object selector and histogram classifier, respectively, inside the boundaries of the television image of the object;
S _OO (n) is the current area of the region inside the boundaries of the television image of the object;
F _{nu_DIF} (n), F _{nu_SDO} (n), F _{nu_GK} (n) - functions for entering the initial conditions of the detector of changes in the background, the selector of moving objects and the histogram classifier;
n is the number of the current frame.

NX, NY - размеры текущего окна анализа;
W_ГК(n), W_ДИФ(n), W_СДО(n) - коэффициенты достоверности сигналов бинарных телевизионных изображений гистограммного классификатора, сигналов вторичных бинарных телевизионных изображений детектора изменений фона и селектора движущихся объектов соответственно;
G_{пр ГК}(iх) и V_{пр ГК}(jу), G_{пр ДИФ}(iх) и V_{пр ДИФ}(jу), G_{пp СДО}(ix) и V_{пр СДО}(jу) - горизонтальные и вертикальные проекции сигналов бинарных телевизионных изображений гистограммного классификатора, сигналов вторичных бинарных телевизионных изображений детектора изменений фона и селектора движущихся объектов соответственно;
n - номер текущего кадра.The technical result is also achieved by the fact that the generalized horizontal and vertical projections _Gpro (ix, n), _Vpro (jy, n) of the signals of the generalized binary television image of the object are formed by a weighted summation of the projections of the binary television images of the background change detector, moving object selector and histogram classifier :
G _{pr O} (ix, n) = W _DIF (n) * G _{pr DIF} (ix) + W _LMS (n) * G _{pr LSS} (ix) + W _GK (n) * G _{pr GK} (ix),
V _{pr O} (jy, n) = W _DIF (n) * V _{ave DIF} (jy) + W _LMS (n) * V _{direct LMS} (jy) + W _CC (n) * V _{ave HA} (jy),
where ix, jy are the coordinates of the television image signals relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
W _ГК (n), W _ДИФ (n), W _ССО (n) - reliability coefficients of signals of binary television images of a histogram classifier, signals of secondary binary television images of a background change detector and a moving object selector, respectively;
G _{ave HA} (ix) and V _{ave HA} (ju), G _{etc. DIF} (ix) and V _{ave DIF} (ju), G _{prosp LMS} (ix) and V _{ave LMS} (ju) - horizontal and vertical projection signals binary television pictures a histogram classifier, signals of secondary binary television images of the background change detector and the moving object selector, respectively;
n is the number of the current frame.

The technical result is also achieved by the fact that the current coordinates X _OBIN , Y _{OBIN of the} generalized binary television image of the object is determined as the weighted sum of the coordinates of the center of gravity X _OTC , Y _OCT and the coordinates X _OMED , Y _{OMED of the} median area of the generalized binary television image of the object
X _OBIN (n) = W _CT (n) * X _OSTT + W _MED (n) * X _OMED ,
Y _OBIN (n) = W _CT (n) * Y _OCT + W _MED (n) * Y _OMED ,
W _MED (n) = 1-W _CT (n),
where W _MED (n), W _CT (n) are the weighting coefficients of the coordinates of the median of the area and the center of gravity of the generalized binary television image of the object, respectively;
n is the number of the current frame,
moreover, the weight coefficient W _CT (n) is increased with a decrease in the average deviation of the coordinates of the television image of the object from their predicted values.

The technical result is also achieved by the fact that the current horizontal VG _{OB BIN} and vertical VV _{OB BIN} components of the estimation of the speed of movement of a binary television image of an object in an inertial coordinate system are determined in accordance with the expressions
VG _{OB_BIN} = (dX + ΔX _OA + ΔX _{OB_OA_BIN} ) / T,
VV _{OB_BIN} = (dY + ΔY _OA + ΔY _{OB_OA_BIN} ) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OB_OA_BIN} , ΔY _{OB_OA_BIN} - change the coordinates of the binary television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively.

NX, NY - размеры текущего окна анализа,
F_НФ[L] - функция двухстороннего ограничения;
F_НФ[L] = L_ПОР при L > L_ПОР;
F_НФ[L] = L при -L_ПОР ≤ L <L_ПОР;
F_НФ[L] = -L_ПОР при L <-L_ПОР;
L_ПОР - пороговый уровень функции двухстороннего ограничения;

К_НФ, К_ФВЧ - постоянные коэффициенты;
di,dj - внутренние переменные фильтра верхних частот,
NF, MF - параметры апертуры фильтра верхних частотно горизонтали и вертикали.The technical result is also achieved by the fact that non-linear high-frequency filtering of signals of a scaled television image in the current analysis window is performed in accordance with the expression
L _nNF (ix, jy) = L _nOA (ix, jy) + K _NF * F _NF [L _nFVF (ix, jy)],
where ix, jy are the coordinates of the television image signals relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
F _NF [L] - function of two-way restriction;
F _NF [L] = L _POR for L> L _POR ;
F _NF [L] = L at -L _POR ≤ L <L _POR ;
F _NF [L] = -L _POR for L <-L _POR ;
L _POR - threshold level of the function of two-way restriction;

K _NF , K _HPF - constant coefficients;
di, dj are the internal variables of the high-pass filter,
NF, MF - filter aperture parameters of the upper frequency horizontal and vertical.

The technical result is also achieved by the fact that the signals of the dynamic reference television image of the object are formed by reading in each frame the signals from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with center, X coordinate _Whose Y _{ET ET} is determined by the difference
X _{C ET} = X _{ABOUT PZ} - X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} - Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view of the video camera of the surveillance system;
X _{OA PZ} , Y _{OA PZ} - coordinates of the center of the analysis window in the field of view of the video camera of the surveillance system.

The technical result is also achieved by the fact that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , Y _{C ET} which is determined by the difference
X _{C ET} = X _{ABOUT PZ} - X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} - Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view of the video camera of the surveillance system;
X _{OA PZ} , Y _{OA PZ} - coordinates of the center of the analysis window in the field of view of the video camera of the surveillance system,
when the conditions for changing the static reference television image of the object formed by comparing the parameters of the signals of the measures of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static and dynamic reference television images of the object, as well as comparing the parameters of the trajectories of the television image of the object obtained when using the signals static and dynamic reference television images of the object.

The technical result is also achieved by the fact that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , U _{C ET} which is determined by the difference
X _{C ET} = X _{ABOUT PZ} - X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} - Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view of the video camera of the surveillance system;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system,
when the conditions for changing the static reference television image of the object, formed on the basis of the analysis of the signal parameters of the measure of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static reference television image of the object, as well as the path parameters of the television image of the object obtained on the basis of signal analysis dissimilarity measures.

The technical result is also achieved by the fact that the current horizontal VG _{OB_NXX} and vertical VV _{OB_NXX} components of the estimation of the speed of movement of the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, are determined in accordance with the expressions
_{OB_HCX} VG = (dX + _OA + ΔH ΔH _{OB_OA NSH)} / T,
_{OB_NSKH} VV = (dY + ΔY _{OA OB_OA NSH} + ΔY) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OB_OA_NXX} , ΔY _{OB_OA_NXX} - change the coordinates of the television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, obtained on the basis of the formation of signals of the measure of dissimilarity of television images.

обобщенного бинарного телевизионного изображения, ограничения минимальных и максимальных значений текущей плотности V_БИН(n) обобщенного бинарного телевизионного изображения, последующего усреднения ограниченной плотности обобщенного бинарного телевизионного изображения рекурсивным фильтром первого порядка, нормирования усредненной плотности

обобщенного бинарного телевизионного изображения объекта

где S_ОБИН(n) - текущая площадь обобщенного бинарного телевизионного изображения внутри границ обобщенного бинарного телевизионного изображения объекта;
S_OO(n) - текущая площадь области внутри границ обобщенного бинарного телевизионного изображения объекта;
n - номер текущего кадра;

усредненный коэффициент сходства, получаемый в результате определения текущего коэффициента сходства

ограничения его максимальных и минимальных значений и усреднения рекурсивным фильтром первого порядка;
σ_фmin(n) - минимальное значение среднеквадратического значения сигналов изображения телевизионного фона в М окнах анализа фона;
Е_НСХmin(n) - минимальное значение сигналов меры несходства телевизионного изображения в текущем окне анализа и статического эталонного телевизионного изображения объекта в двумерной области поиска смещений телевизионного изображения объекта.The technical result is also achieved by the fact that the confidence coefficient W _BIN (n) of the current velocity of the generalized binary television image of the object is obtained by determining the current density

generalized binary television image, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image, then averaging the limited density of the generalized binary television image with a first order recursive filter, normalizing the average density

generalized binary television image of an object

where S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object;
S _OO (n) is the current area of the region inside the boundaries of the generalized binary television image of the object;
n is the number of the current frame;

average similarity coefficient obtained by determining the current similarity coefficient

restrictions on its maximum and minimum values and averaging by a first order recursive filter;
σ _fmin (n) - the minimum value of the rms value of the image signals of the television background in M windows analysis background;
Е _НСХmin (n) - the minimum value of the signals of the measure of dissimilarity of the television image in the current analysis window and the static reference television image of the object in the two-dimensional region for searching for the displacements of the television image of the object.

ограничения его максимальных и минимальных значений и усреднения рекурсивным фильтром первого порядка, нормирования полученного усредненного коэффициента сходства

где n - номер текущего кадра;
σ_фmin(n) - - минимальное значение среднеквадратического значения сигналов телевизионного изображения фона в М окнах анализа фона;
Е_НСХmin(n) - минимальное значение сигналов меры несходства телевизионных изображений текущего окна анализа и статического эталонного телевизионного изображения объекта в двумерной области поиска смещений телевизионного изображения объекта;

усредненная плотность обобщенного бинарного телевизионного изображения объекта, получаемая в результате определения текущей плотности

обобщенного бинарного телевизионного изображения объекта, ограничения минимальных и максимальных значений текущей плотности V_БИН(n) обобщенного бинарного телевизионного изображения объекта и последующего усреднения рекурсивным фильтром первого порядка ограниченной текущей плотности V_БИН(n) обобщенного бинарного телевизионного изображения объекта;
S_{О БИН}(n) - текущая площадь обобщенного бинарного телевизионного изображения внутри границ обобщенного бинарного телевизионного изображения объекта,
S_ОО(n) - текущая площадь области внутри границ обобщенного бинарного телевизионного изображения объекта.The technical result is also achieved by the fact that the confidence coefficient W _CX (n) of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, is obtained by determining the current coefficient of similarity

limiting its maximum and minimum values and averaging with a first-order recursive filter, normalizing the obtained average similarity coefficient

where n is the number of the current frame;
σ _fmin (n) - is the minimum value of the rms value of the background television image signals in M background analysis windows;
Е _НСХmin (n) - the minimum value of the signals of the measure of dissimilarity of the television images of the current analysis window and the static reference television image of the object in the two-dimensional region for searching for the displacements of the television image of the object;

averaged density of a generalized binary television image of an object obtained by determining the current density

generalized binary television image of the object, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image of the object and then averaging, with a first order recursive filter, the limited current density V _BIN (n) of the generalized binary television image of the object;
S _{O BIN} (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object,
S _ОО (n) is the current area of the region inside the boundaries of the generalized binary television image of the object.

The technical result is also achieved by the fact that a comprehensive assessment of the horizontal VG _OB (n) and vertical VV _OB (n) components of the current speed of the television image of the object in the inertial coordinate system is determined by limiting the estimates of the speed of the binary television image of the object and the speed of the television image of the object obtained on the basis of the formation of signals of a measure of dissimilarity of television images, the minimum and maximum values formed taking into account pre striding values complex evaluation of the current speed of movement of a television image of the object, and forming a weighted sum of the count rate of movement limited television binary object image and the speed of movement of television images of the object obtained based on the signal generating dissimilarity measure television pictures
VG _ON (n) = W _BIN (n) * VG _{OB_BIN} (n) + W _SC (n) * VG _{HSKH ON} (n),
VV _ON (n) = W _BIN (n) * VV _{OB_BIN} (n) + W _SC (n) * VV _{HSKH ON} (n),
where W _BIN (n), W _CX (n) are the reliability coefficients of the current velocity of the generalized binary television image of the object and the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, respectively;
VG _{OB_BIN} (n) and VV _{OB_BIN} (n) - limited horizontal and vertical components of the current speed of movement of the generalized binary television image of the object;
VG _{ABOUT NSX} (n) and VV _{ABOUT NSX} (n) - limited horizontal and vertical components of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images;
n is the number of the current frame.

площади обобщенного бинарного телевизионного изображения объекта производят проверкой выполнения условия

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта или условия,

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа,
где ksl - постоянный коэффициент, ksl < 1;
ks2(n-n_ЭК) - коэффициент, уменьшающийся с ростом номера кадра n, начиная с номера кадра n_ЭК, перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам, ks2(n-n_ЭК)≤ksl.The technical result is also achieved by the fact that the analysis of the current S _OBIN (n) and averaged

the area of the generalized binary television image of the object is produced by checking the fulfillment of the condition

to switch to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the extrapolated coordinates of the television image of the object or condition,

to proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window,
where ksl is a constant coefficient, ksl <1;
ks2 (nn _EC ) is a coefficient that decreases with increasing frame number n, starting from frame number n _EC , the transition to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates, ks2 (nn _EC ) ≤ksl.

- для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам или выполнения условий

- для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа,
где VG_TОБ(n), VV_TOБ(n) - горизонтальная и вертикальная составляющие текущей скорости перемещения телевизионного изображения объекта, получаемые в результате рекурсивной фильтрации первого порядка с постоянной фильтра 0<WV1<1 составляющих VG_ОБ(n), VV_ОБ(n) комплексной оценки скорости перемещения телевизионного изображения объекта:
VG_ТОБ(n) = VG_ТОБ(n-1)+ WV1*[VG_ОБ(n)-VG_ТОБ(n-1)],
VV_ТОБ(n) = VV_ТОБ(n-1)+ WV1*[VV_ОБ(n)- VV_ТОБ(n-1)],

горизонтальная и вертикальная составляющие усредненной скорости перемещения телевизионного изображения объекта, получаемые в результате рекурсивной фильтрации первого порядка с постоянной фильтра 0<WV<2 составляющих VG_ОБ(n), VV_ОБ(n) комплексной оценки скорости перемещения телевизионного изображения объекта:

причем WV1>WV2.The technical result is also achieved by the fact that the analysis of the current and average speed of the television image of the object is carried out by checking the conditions

- to switch to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates or to fulfill the conditions

- to proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window,
where VG _TOB (n), VV _TOB (n) are the horizontal and vertical components of the current speed of the television image of the object, obtained as a result of first-order recursive filtering with a filter constant of 0 <WV1 <1 components of VG _OB (n), VV _OB (n ) a comprehensive assessment of the speed of movement of the television image of the object:
VG _TOB (n) = VG _TOB (n-1) + WV1 * [VG _ON (n) -VG _TOB (n-1)],
_TOB VV (n) = VV _TOB (n-1) + WV1 * [VV _ON (n) - VV _TOB (n-1)],

the horizontal and vertical components of the average speed of moving the television image of the object, obtained as a result of first-order recursive filtering with a filter constant of 0 <WV <2 components VG _OB (n), VV _OB (n) of a complex estimate of the speed of the television image of the object:

moreover, WV1> WV2.

где S_i(n) - значения текущей площади бинарных телевизионных изображений гистограммного классификатора внутри i-гo окна анализа фона, i = 1,...,М;

значения усредненной площади бинарных телевизионных изображений гистограммного классификатора внутри i-гo окна анализа фона, i = 1,...,М, получаемые на выходе рекурсивного фильтра первого порядка с постоянной фильтра 0<WS<1:

Sпорог(n) - порог обнаружения изменения площади бинарных телевизионных изображений гистограммного классификатора внутри М окон анализа фона:

где k - постоянный коэффициент;

усредненная площадь обобщенного бинарного телевизионного изображения объекта, получаемая на выходе рекурсивного фильтра первого порядка с постоянной фильтра 0<WS_ОБ<1:

n - номер текущего кадра;

- отношение объект/фон,
где

среднее значение сигналов телевизионного изображения в окне объекта;

среднее значение сигналов телевизионного изображения в окне фона;
σ_Ф - среднеквадратическое отклонение сигналов телевизионного изображения в окне фона от

F_ОГР(Q) - функция ограничения минимальных и максимальных значений;
при выполнении указанных выше условий для текущей площади бинарных телевизионных изображений гистограммного классификатора в одном или нескольких окнах анализа фона с номерами ki,
где ki - номера окон анализа фона, в которых обнаружено изменение текущей площади бинарного телевизионного изображения гистограммного классификатора, ki=l,...,кN;
кN - число окон анализа фона, в которых обнаружено изменение текущей площади бинарного телевизионного изображения гистограммного классификатора, кN ≤ М;
в этих окнах анализируют координаты двух взаимно перпендикулярных границ бинарного телевизионного изображения гистограммного классификатора, расположенных со стороны окна объекта, причем в окнах анализа фона, размещенных со стороны вертикальных границ окна объекта, анализируют перемещение вертикальной границы бинарного телевизионного изображения по горизонтали и контролируют принадлежность координат горизонтальной границы вертикальным координатам окна объекта, а в окнах анализа фона, размещенных со стороны горизонтальных границ окна объекта, анализируют перемещение горизонтальной границы бинарного телевизионного изображения по вертикали и контролируют принадлежность координат вертикальной границы горизонтальным координатам окна объекта, при этом анализ перемещения границ бинарных телевизионных изображений в М окнах анализа фона проводят посредством проверки условий нахождения соответствующих границ бинарного телевизионного изображения гистограммного классификатора внутри границ внутренней O_kil и внешней O_ki2 областей ki-го окна анализа фона с формированием признаков П_ki1 и П_ki2 принадлежности границ бинарных телевизионных изображений к внутренней О_ki1 и к внешней O_ki2 областям ki-го окна анализа фона,
где П_ki1= 1, если граница бинарного телевизионного изображения гистограммного классификатора лежит внутри границ области О_ki1, или П_ki1=0 - в противном случае;
П_ki2=1, если граница бинарного телевизионного изображения гистограммного классификатора лежит внутри границ области O_ki2, или П_ki2=0 - в противном случае;
причем при обнаружении переходов соответствующих границ бинарного телевизионного изображения гистограммного классификатора от внешней области к внутренней области окон анализа фона с номерами ki=kipl,
где kipl - номера окон анализа фона, в которых обнаружены переходы границ бинарного телевизионного изображения гистограммного классификатора из внешних областей O_kip12 к внутренним областям O_kip11 окон анализа фона, kipl=l,2,...,kN,
то есть при последовательном формировании признаков П_kip12=l при n_kip1= n_kip12, а затем П_kip1=1 при n_kip1=n_kip11, где n_kip11>n_kip12, n_kip11 - номер кадра, в котором в kip1-м окне анализа фона установлено состояние признака П_kip11= 1, n_kip12 - номер кадра, в котором в kipl-м окне анализа фона установлено состояние признака П_kip12=1, и выполнении условия принадлежности координаты второй контролируемой границы бинарного телевизионного изображения гистограммного классификатора в kipl-м окне анализа фона диапазону координат окна объекта, выполняют переход к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта,
присваивают счетчику SP переходов границ бинарного телевизионного изображения гистограммного классификатора между внутренними и внешними областями окон анализа фона значение SP= 1 при первом обнаружении перехода границы бинарного телевизионного изображения гистограммного классификатора из внешней области O_kip12 к внутренней области O_kip11 или увеличивают счетчик SP на единицу при повторном обнаружении переходов границы бинарного телевизионного изображения гистограммного классификатора из внешних областей O_kip12 к внутренним областям O_kip11 в течение формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта,
устанавливают состояния признаков П_kip11=0, П_kip12=0, устанавливают состояния признаков П_j2=0 в окнах анализа фона с номерами j≠kip1, j=1,...,кN, и процесс анализа границ бинарного телевизионного изображения гистограммного классификатора начинают сначала,
при обнаружении переходов соответствующих границ бинарного телевизионного изображения гистограммного классификатора из внутренних областей O_kip21 к внешним областям O_kip22 окон анализа фона,
где kiр2 - номера окон анализа фона, в которых обнаружен переход границы бинарного телевизионного изображения гистограммного классификатора из внутренних областей O_kip22 к внешних областям O_kip21 окон анализа фона, kip2=l, 2,...,кN,
в течение формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта, то есть при последовательном формировании признаков П_kiр21=1 при n_kip2=n_kip21, а затем П_kiр22=1 при n_kiр2=n_kiр22 в одном или нескольких окнах анализа фона с номерами kip2, где n_kiр22>n_kiр21, уменьшают счетчик SP-переходов границ бинарного телевизионного изображения гистограммного классификатора между внутренними и внешними областями окон анализа фона на единицу и, если SP=0, выполняют переход к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа, устанавливают состояния признаков П_kip21=0, П_kip22=0, устанавливают состояния признаков П_j2= 0 в окнах анализа фона с номерами j≠kip2, j=i,...,кN, и процесс анализа границ бинарного телевизионного изображения гистограммного классификатора начинают сначала, контролируют время формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам и время нахождения признаков П_i1, П_i2, i=l,...,M, в состоянии П_i1=1, П_i2=1 и при превышении установленных интервалов времени переводят соответствующие признаки в нулевое состояние, переходят к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа.The technical result is also achieved by the fact that the analysis of the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows analysis of the background is carried out by checking the conditions for detecting changes in the area of the binary television images of the histogram classifier inside the M windows of the background analysis:

where S _i (n) are the values of the current area of binary television images of the histogram classifier inside the i-th window of the background analysis, i = 1, ..., M;

values of the average area of binary television images of the histogram classifier inside the ith window of the background analysis, i = 1, ..., M, obtained at the output of the first order recursive filter with a filter constant of 0 <WS <1:

S threshold (n) - threshold for detecting a change in the area of binary television images of a histogram classifier inside M windows of background analysis:

where k is a constant coefficient;

the average area of the generalized binary television image of the object obtained at the output of a first-order recursive filter with a filter constant of 0 <WS _OB <1:

n is the number of the current frame;

- object / background ratio,
Where

the average value of the television image signals in the object window;

the average value of the television image signals in the background window;
σ _f - the standard deviation of the signals of the television image in the background window from

F _OGR (Q) - the function of limiting the minimum and maximum values;
when the above conditions are met for the current area of binary television images of the histogram classifier in one or more background analysis windows with numbers ki,
where ki are the numbers of the background analysis windows in which a change in the current area of the binary television image of the histogram classifier was detected, ki = l, ..., kN;
kN is the number of background analysis windows in which a change in the current area of the binary television image of the histogram classifier was detected, kN ≤ M;
in these windows, the coordinates of two mutually perpendicular boundaries of the binary television image of the histogram classifier located on the side of the object window are analyzed, and in the background analysis windows located on the side of the vertical borders of the object window, the horizontal movement of the vertical border of the binary television image is analyzed and the horizontal coordinate coordinates are controlled the vertical coordinates of the object’s window, and in the background analysis windows placed on the side of horizontal faces the object’s windows, analyze the vertical horizontal displacement of the binary television image and control the vertical border coordinates of the horizontal coordinates of the object’s window, while the analysis of the displacement of the binary television image boundaries in the M analysis windows is carried out by checking the conditions for finding the corresponding boundaries of the histogram binary television image within the boundaries internal O _ki l and external O _ki 2 areas of the ki-th background analysis window with knowledge of the signs P _ki 1 and P _ki 2 that the boundaries of the binary television images belong to the internal O _ki 1 and to the external O _ki 2 regions of the ki-th background analysis window
where П _ki 1 = 1, if the boundary of the binary television image of the histogram classifier lies inside the boundaries of the region О _ki 1, or П _ki 1 = 0 otherwise;
П _ki 2 = 1, if the boundary of the binary television image of the histogram classifier lies within the boundaries of the region O _ki 2, or П _ki 2 = 0 otherwise;
moreover, when detecting transitions of the corresponding boundaries of the binary television image of the histogram classifier from the outer region to the inner region of the background analysis windows with numbers ki = kipl,
where kipl are the numbers of the background analysis windows in which transitions of the boundaries of the binary television image of the histogram classifier from the outer regions O _kip1 2 to the inner regions O _kip1 1 of the background analysis windows, kipl = l, 2, ..., kN,
that is, in the sequential formation of signs П _kip1 2 = l for n _kip 1 = n _kip1 2, and then П _kip 1 = 1 for n _kip1 = n _kip1 1, where n _kip1 1> n _kip1 2, n _kip1 1 is the frame number in which the state of the attribute П _kip1 1 = 1, n _kip1 2 is set in the kip1th window of the background analysis, n _kip1 2 is the frame number in which the state of the sign П _kip1 2 = 1 is set in the kipl-window of the background analysis, and the second coordinate condition is satisfied of the controlled boundary of the binary television image of the histogram classifier in the kipl-m window of the background analysis to the coordinate range of the window of the object, the transition to the formation of als axis motion control system camcorder view field observation extrapolated coordinates of the television picture of the object,
set the counter SP of transitions of the boundaries of the binary television image of the histogram classifier between the internal and external areas of the background analysis windows to SP = 1 upon the first detection of the transition of the boundary of the binary television image of the histogram classifier from the outer region O _kip1 2 to the inner region O _kip1 1 or increase the counter SP by one upon repeated detection of transitions of the boundary of the binary television image of the histogram classifier from the outer regions O _kip1 2 to the inner regions O _{ki p1} 1 during the generation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the extrapolated coordinates of the television image of the object,
establish the states of signs П _kip1 1 = 0, П _kip1 2 = 0, set the states of signs П _j 2 = 0 in the background analysis windows with numbers j ≠ kip1, j = 1, ..., кN, and the process of analyzing the boundaries of the binary television image the histogram classifier start over,
upon detecting transitions of the corresponding boundaries of the binary television image of the histogram classifier from the inner regions O _kip2 1 to the outer regions O _kip2 2 of the background analysis windows,
where kip2 are the numbers of the background analysis windows in which the transition of the boundary of the binary television image of the histogram classifier from the inner regions O _kip2 2 to the outer regions O _kip2 1 of the background analysis windows, kip2 = l, 2, ..., kN,
during the generation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by the extrapolated coordinates of the television image of the object, that is, with the sequential formation of signs P _kip2 1 = 1 for n _kip2 = n _kip2 1 and then P _kip2 2 = 1 for n _kip2 = n _kip2 2 in one or several background analysis windows with numbers kip2, where n _kip2 2> n _kip2 1, reduce the counter of SP transitions of the boundaries of the binary television image of the histogram classifier between the internal and external regions of the background analysis windows by one and, if SP = 0, weeding they take the transition to the formation of control signals for moving the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window, set the status of signs П _kip2 1 = 0, П _kip2 2 = 0 is set state signs n _j 2 = 0 in the analysis windows background numbers kip2 ≠ j, j = i, ..., kN, and the process boundary analysis binary television image histogram classifier begin first, to ntroliruyut formation time axis motion control signals video camera field of view for observation system extrapolated coordinates and time finding signs n _i 1, n _i 2, i = l, ..., M , are able P _i 1 = 1, P 2 _i = 1 and when the specified time intervals are exceeded, the corresponding signs are converted to the zero state, proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained in p television image processing result in the current analysis window.

The technical result in terms of providing coordinates and stable image retention of the object in the center of the field of view of the video camera of the monitoring system or in the center of the tracking window when the camera receives both contrast and low-contrast images of moving and stationary objects and the background of the terrain, when the object moves along the terrain, the average brightness of which coincides with the average brightness of the object, when changing the visible image of the object, due to its rotation during movement, change in illumination, obs By erecting the object in conditions of intermittent optical communication arising due to the obstruction of the object by various obstacles, such as folds of the terrain, buildings, vegetation, dust, smoke, splashes, etc. P. , with a controlled (mobile) and uncontrolled (fixed) field of view of the video camera, in the presence of motion of the carrier of the surveillance system, causing low-frequency uncontrolled movements of the field of view of the video camera of the surveillance system (which lead to displacements and rotations of all elements of the field of the television image as a whole), and with the possible absence in the monitoring system of sensors of errors of stabilization of the field of view of the video camera, the monitoring system is achieved by the fact that in the device containing the signal generation unit (4) a measure of the dissimilarity of television images of 2N landmarks and the formation of shift and rotation parameters of the signals of the current field of the television image between the reception of signals of the current and previous fields of television images and the processor (24) for processing local data transmitted via a common bi-directional bus, a block (1 ) receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device, unit (2) p the method and storing of signals of uncontrolled movement and roll of the field of view of the video camera of the monitoring system, unit (3) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of the signals of the (nl) -th and (n-2) -th fields of the television image, where n = 3, 4, 5 ,. . . , is the number of the current field of the television image, the separation unit (5) obtained on the basis of the measure of dissimilarity of the television images 2N landmarks of the shift parameters of the signals of the current field of the television image during the time between receiving signals of the current and previous fields of television images on the components of the controlled movement of the axis of the field of view of the video camera surveillance system and uncontrolled movement of the axis of the field of view of the video camera surveillance system, switch block (6), signal shaper (7) of the visual image of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals of the (n-1) th and (n-2) th fields of the television image and the signals of the current field a television image with an element-wise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, the unit (8) of formation and m scaling of the television image signals in the current analysis window using the position and size signals of the current television image analysis window, block (9) for generating M background analysis windows along the perimeter of the analysis window and determining the projections of the binary television image signals of the histogram classifier in M background analysis windows, block ( 10) determining the current coordinates of a generalized binary television image of an object using generalized horizontal and vertical projections in general about a binary television image of an object, a block (11) for determining the coordinates of a television image of an object relative to the center of the current analysis window based on the generation of signals of a measure of dissimilarity of television images, a block (12) for determining the current speed of a television image of an object in an inertial coordinate system using the coordinates of a television image of an object obtained on the basis of the formation of signals of a measure of dissimilarity of television images, the switch (13) source codes or t the current dimensions of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, a unit (14) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in the inertial coordinate system, block (15) for determining the reliability coefficients of the current speed of the television image of the object in the inertial coordinate system obtained on the basis of the formation of signals of a measure of dissimilarity of television images, block (16) is determined determining the area and coordinates of the boundaries of the binary television image in the M windows of analysis of the background, the block (17) for determining the current speed of the generalized binary television image of the object in the inertial coordinate system, the block (18) for generating a complex estimate of the current speed of the television image of the object in the inertial coordinate system, block (19) for an integrated estimation of the coordinates of a television image of an object in the field of view of a video camera of a surveillance system, an analyzer (20) of the conditions for transition to use extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in M background analysis windows, unit (21) for averaging the complex estimate of the current the speed of moving the television image of the object and storing the values of the average speed, block (22) determining prog the coordinates and the speed of the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of the television image of the object, the shaper (23) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of a television image of an object in the field of view of a video camera of a surveillance system or extrapol coordinates and the speed of the television image of the object, and the first and second inputs of the device are connected to the first and second inputs of the unit (1) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating synchronization signals of the device, respectively, the third input of the device is connected to the input of the block (2) receiving and storing signals of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system, the fourth and fifth inputs of the device soy inens with the first and second inputs of the switch (13) codes of the source or current dimensions of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, respectively, the sixth input of the device is connected to the second input of the block (3) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between the reception of signals of the (nl) th and (n-2) th fields of the television image, where n = 3, 4, 5 ,. . . , - the number of the current field of the television image, the seventh input of the device is connected to the fifth input of the block (6) of the switches, the first output of the block (1) of receiving and storing signals of the current field of the television image from the video camera of the system for monitoring and generating signals for synchronizing the operation of the device is connected to the first input of the unit (4) generating signals of a measure of dissimilarity of television images of 2N landmarks and generating shift and rotation parameters of signals of the current field of the television image during the time between signals of the current and previous fields of television images and with the first input of the imager (7) of the television image signals of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals (n-1) -th and (n-2) -th fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the monitoring system, the second output of the block (1) for receiving and storing signals of the current field of the television image from the video camera of the monitoring and generation of synchronization signals of the device is connected to the third input of the block (3) for calculating the controlled movement of the axis of the camera’s field of view monitoring systems for the time between reception of signals of the (n-1) -th and (n-2) -th fields of a television image, where n = 3, 4, 5 ,. . . , is the number of the current television image field, the fifth input of the television image signal generation and scaling unit (8) in the current analysis window using position and size signals of the current television image analysis window, the fourth input of the television image signal generator (7) of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th lei of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, by the second input of the separation unit (5) obtained on the basis of the measure of dissimilarity of television images 2N landmarks of terrain parameters the shift of the signals of the current field of the television image during the time between receiving signals of the current and previous fields th television images on the components of the controlled movement of the axis of the field of view of the video camera of the monitoring system and the uncontrolled movement of the axis of the field of view of the video camera of the monitoring system, the fourth input of the block (4) for generating signals of the measure of dissimilarity of television images 2N landmarks and the formation of the shift and rotation parameters of the signals of the current field of the television image for the time between the reception of signals of the current and previous fields of television images by the second inputs of the block (9) of the formation of M windows ana liza of the background around the window of analysis and determination of projections of signals of a binary television image of a histogram classifier in M windows of analysis of the background and block (16) for determining the area and coordinates of the boundaries of binary television images in M windows of analysis of the background, the third input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth input of the block (11) is determined the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of television images, the fourth input of the block (12) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object obtained on the basis of the formation of signals of the measure of dissimilarity television images, the fifth input of the switch (13) codes of the original or current sizes of the television images of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, the fourth inputs of the block (14) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in the inertial coordinate system, and the block (15) for determining the reliability coefficients of the current speed of the television image of the object in the inertial the coordinate system obtained on the basis of the formation of signals of the measure of dissimilarity of television images, the fourth block input and (17) determining the current velocity of the generalized binary television image of the object in the inertial coordinate system, the fifth input of the block (18) generating a complex estimate of the current speed of the television image of the object in the inertial coordinate system, the third input of the block (19) of the complex estimation of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the sixth input of the analyzer (20) conditions for the transition to the use of extrapolated coordinates of the television from object reflection based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in M background analysis windows, the third inputs of the averaging unit (21) for complex estimation of the current movement speed television image of the object and storing the values of the average speed, and the block (22) determine the predicted coordinates and speed the movement of the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of movement of the television image of the object, and the sixth input of the shaper (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the surveillance system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system or extrapolated co the distance and speed of the television image of the object, the output of the block (2) for receiving and storing the signals of uncontrolled movement and the roll of the field of view of the video camera of the monitoring system is connected to the third input of the block (4) for generating signals of the measure of dissimilarity of the television images 2N of the terrain landmarks and the formation of shift and rotation parameters of the signals the current field of the television image during the time between the reception of signals of the current and previous fields of the television image and with the third input of the block (6) of switches the first output of the block (3) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) th and (n-2) th fields of the television image, where n = 3, 4, 5 ,. . . , - number of the current field of the television image, connected to the second input of the block (4) generating signals of the measure of dissimilarity of television images 2N landmarks and the formation of the parameters of the shift and rotation of the signals of the current field of the television image during the time between receiving signals of the current and previous fields of television images and to the fourth input of the block (6) of switches the first output of the block (4) for generating signals of a measure of dissimilarity of television images of 2N landmarks and the formation of parameters for shifting and rotating the signals of the current field of the television image during the time between receiving signals from the current and previous fields of television images and connected to the first input of the separation block (5) obtained on the basis of estimates of the dissimilarity measure of television images of 2N landmarks of the shift parameters of the signals of the current field of the television image during the time between the reception of signals x and previous fields of television images on the components of the controlled movement of the axis of the field of view of the video camera of the surveillance system and the uncontrolled movement of the axis of the field of view of the video camera of the surveillance system, the first and second outputs of which are connected to the first and second inputs of the block (6) of the switches, the first output of the block (6) of switches is connected to the second input of the shaper (7) of the television image signals of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals (n-1) - th and (n-2) th fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movement and roll of the field of view of the video camera of the surveillance system, and the second output is connected to the third input of the shaper (7) of the television image signals of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n -2) th fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, with the fifth input of the block (10) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the first input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, with the first input of the block (12) determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, with the second input of the block (17) determining the current speed of the generalized binary television image of the object in an inertial coordinate system, and with the first input of the block (19) of an integrated assessment of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the output of the shaper (7) of the television image signals of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, connected to the first input of a block (8) for generating and scaling television image signals in the current analysis window using position and size signals of the current television image analysis window, the output of the block (8) for generating and scaling television image signals in the current analysis window using the position and size signals of the current television image analysis window is connected to the seventh input of the block (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of the generalized binary television image of the object and with the third input of the block (11) determining the coordinates of the television image of the object relative to the current analysis window center on the basis of a signal generating dissimilarity measure television images, the output of the block (9) for generating M windows for analyzing the background around the perimeter of the window for analyzing and projecting the signals of the binary television image of the histogram classifier in the M windows for analyzing the background is connected to the first input of the block (16) for determining the area and coordinates of the boundaries of the binary television images in the M windows for analyzing the background, the first output of which is connected to the third input of the analyzer (20) of the transition conditions to use the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the first output of the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the block (9) for the formation of M background analysis windows along the perimeter of the analysis window and for determining the projections of the signals of the binary television image histogram classifier in M background analysis windows, the second output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the switch (13) of the codes for the initial or current sizes of the television image of the object and the coordinates of the object in the field of view of the system’s video camera observations the first input of the block (14) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in the inertial coordinate system and the first input of the shaper (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image with using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system or extrapolated coordinates and awns moving television image of the object, the third output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the block (17) for determining the current speed of the movement of the generalized binary television image of the object in an inertial coordinate system, the fourth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the second input of the block (14) for determining the reliability coefficients of the current speed of movement of the generalized binary television image of the object in the inertial coordinate system and with the second input of the analyzer (20), the conditions for the transition to the use of extrapolated coordinates evizionnogo image of the object based on the current and the average area of a generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the fifth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the analyzer (20) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of an object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the sixth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the block (15) for determining the reliability coefficients of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the seventh output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the fourth input of the block (1) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating synchronization signals device operation the first output of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the second input of the block (12) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the third input of the block (18) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the second output of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the first input of the block (15) for determining the reliability coefficients of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the second input of the block (18) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the first output of the switch (13) of codes of the source or current dimensions of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the first input of the block (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of the generalized binary television image facility the seventh input of the unit (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the third input of the unit (8) for generating and scaling the signals of the television image in the current analysis window using the position and size signals of the current analysis window television image the second output of the switch (13) of codes of the initial or current dimensions of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the second input of the block (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of the generalized binary television image facility the sixth input of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the second input of the block (8) for generating and scaling the signals of the television image in the current analysis window using the position and size signals of the current analysis window television image the first output of the block (14) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in the inertial coordinate system is connected to the first input of the block (18) for forming a complex estimate of the current speed of the television image of the object in the inertial coordinate system, and the second output is connected to the second input of the block (15) for determining the reliability coefficients of the current speed of movement of the television image of the object in an inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the second output of which is connected to the third input of the block (14) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in an inertial coordinate system, the first output of the block (17) for determining the current moving speed of the generalized binary television image of the object in the inertial coordinate system is connected to the fourth input of the block (18) for generating a complex estimate of the current speed of moving the television image of the object in the inertial coordinate system, the first output of the unit (18) for generating a comprehensive estimate of the current speed of moving the television image of the object in an inertial coordinate system is connected to the fourth input of the analyzer (20) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the fifth input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, the first input of the unit (21) for averaging a complex estimate of the current speed of moving a television image of an object and storing the values of the average speed and the second input of a block (19) for a complex estimate of the coordinates of a television image of an object in the field of view of a video camera of a surveillance system, the first output of the unit (19) for the complex assessment of the coordinates of the television image of the object in the field of view of the video camera of the monitoring system is connected to the fourth input of the switch (13) codes of the initial or current sizes of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, the first input of the unit (22) for determining the predicted coordinates and the speed of the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of the television image of the object and the third input of the driver (23) of the control signals for moving the axis of the field of view of the video camera of the observation system and position signals and dimensions of the analysis window in the next frame of the television image using the coordinates of the television image of the object in Beyond the view of the video camera of the surveillance system or extrapolated coordinates and the speed of the television image of the object the second output of which is connected to the third inputs of the block (12) for determining the current speed of the television image of the object in an inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of the formation of the signals of the measure of dissimilarity of the television images and the block (17) for determining the current speed of movement of the generalized binary television image of the object in an inertial coordinate system, the sixth input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object and the fourth inputs of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of television dissimilarity images and block (8) forming and scaling
television image signals in the current analysis window using position and size signals of the current television image analysis window, the first output of the analyzer (20) conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, current and average speed moving a television image of an object, the area and coordinates of the boundaries of binary television histogram images classifier in M windows of analysis of the background is connected to the second input of the shaper (23) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and speed of the television image of the object, the second input of the block (21) averaging a comprehensive assessment of the current speed of movement a television image of the object and storing the values of the average speed and with the fourth input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the second output of the analyzer (20) conditions for the transition to the use of extrapolated coordinates a television image of an object based on the current and average area of a generalized binary television and the image of the object, the current and average speed of moving the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows of analysis of the background is connected to the second input of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of dissimilarity measures television images, the fourth output of the analyzer (20) conditions for the transition to the use of extrapolated television coordinates image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in M background analysis windows is connected to the third input of the block (1) for receiving and storing signals of the current fields of a television image from a video camera of the monitoring system and generating signals for synchronizing the operation of the device, the first output of the block (21) is averaged a complex assessment of the current speed of moving the television image of the object and storing the values of the average speed is connected to the fifth input of the analyzer (20) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of movement of the television image of the object, area and coordinates of the boundaries of binary television images of a histogram classification ikator in M windows for analysis of the background and the second input of the block (22) for determining the predicted coordinates and speed of moving the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of moving the television image of the object, the second output of the block (22) for determining the predicted coordinates and speed moving the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of movement of the television The image of the object is connected to the fifth input of the generator (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image in the field of view of the video camera of the monitoring system or extrapolated coordinates and speed television image of the object, the first output of the block (22) determining the predicted coordinates and speed of movement of the television the image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of movement of the television image of the object is connected to the fourth input of the shaper (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using coordinates television image of an object in the field of view of a video camera of a surveillance system or extrapolated coordinates and speed the movement of the television image of the object, the first output of the driver (23) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image in the field of view of the video camera of the monitoring system or extrapolated coordinates and speed The television image of the object is the output of the device and is connected to the first input of the calculation unit (3) on the movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image, where n = 3, 4, 5, ..., is the number of the current field of the television image, the output of the processor (24) for processing local data transmitted via a common bi-directional bus is connected to the second outputs of the block (3) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n -2) th fields of the television image, where n = 3, 4, 5,. . ., - number of the current field of the television image, block (4) for generating signals of a measure of dissimilarity of television images 2N landmarks and the formation of parameters for the shift and rotation of the signals of the current field of the television image for the time between reception of signals of the current and previous fields of television images, block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television image, block (12) determining those the current speed of moving a television image of an object in an inertial coordinate system using the coordinates of a television image of an object obtained on the basis of generating signals of a measure of dissimilarity of television images, a block (16) for determining the area and coordinates of the boundaries of binary television images in M windows for analyzing the background, a block (18) for generating a complex assessment of the current speed of movement of a television image of an object in an inertial coordinate system, block (19) of an integrated assessment of the coordinates of a body visual image of the object in the field of view of the video camera of the monitoring system, unit (22) for determining the predicted coordinates and speed of the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of movement of the television image of the object, the third outputs of the block (14) for determining the reliability coefficients of the current the moving speed of a generalized binary television image of an object in an inertial coordinate system, block (15) is defined reliability coefficients of the current speed of moving the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, the analyzer (20) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and the average speed of the television image of the object, the area and coordinates of the gran q binary television images of the histogram classifier in M windows for analyzing the background, unit (21) for averaging a complex estimate of the current speed of the television image of the object and storing the values of the average speed and driver (23) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and position and window size signals analysis in the next frame of the television image using the coordinates of the television image of the object in the field of view of the camera system is observed or extrapolated coordinates and speed of the television image of the object, the eighth output of the block (1) receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device.

 The technical result is also achieved by the fact that the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object contains a first video data switch (31), a first buffer random access memory (30), and a second buffer random access memory (41) and a third buffer random access memory (43), a shaper (32) of binary television signals histogram classifier images in the current analysis window, primary binary television image signal shaper (36) for background change detector, secondary binary television image signal shaper (38) for background changes detector, signal shaper (45) for primary binary television image of moving objects selector, shaper (47 ) signals of the secondary binary television image of the selector of moving objects, the first node (42) scaling and shifting the signals of the body of a visual image, a node (44) for generating signals of a differential television image of a selector of moving objects, a shaper (33) of horizontal and vertical projections of signals of a binary television image of a histogram classifier, a shaper (39) of horizontal and vertical projections of signals of a secondary binary television image of a detector of background changes, a shaper ( 48) horizontal and vertical projections of the signals of the secondary binary television image of the moving selector objects, block (35) for determining the object / background ratio and the minimum rms value of the background television image signals in M background analysis windows, a spatial filter (37) of the low frequencies of the signals of the primary binary television image of the background change detector, spatial filter (46) of the low signal frequencies primary binary television image of the moving object selector, block (34) for determining the reliability coefficients of signals of binary television images of a histogram classifier a torus, a block (40) for determining the reliability coefficients of the signals of the secondary binary television images of the background change detector, a block (49) for determining the reliability coefficients of the signals of the secondary binary television images of the selector of moving objects and a generator (50) of generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, block (51) for determining coordinates and sizes, current and average area of a generalized binary television image object and in the analysis window, and the block (52) for analyzing the breakdown conditions for automatically determining the coordinates of the television image of the object, the first and second inputs of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object are connected with the first and second inputs of the shaper (32) of the binary television image signals of the histogram classifier in the current analysis and block window (35) determining the object / background ratio and the minimum rms value of the background television image signals in M background analysis windows and the fourth and third inputs of the primary binary television image signal detector for changing the background changes, respectively, the first input of the unit (10) for determining the current coordinates of the generalized a binary television image of an object using generalized horizontal and vertical projections of a generalized binary television image of an object is connected t also with the first input of the first node (42) for scaling and shifting the television image signals, the third input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the first switch (31 ) of video data, the seventh input of the shaper (32) of the binary television image signals of the histogram classifier in the current analysis window, the second the input of the shaper (33) of horizontal and vertical projections of the signals of the binary television image of the histogram classifier, the second input of the block (34) for determining the reliability coefficients of the signals of the binary television images of the histogram classifier, the fifth input of the block (35) for determining the object / background ratio and the minimum rms value of the signals of the television image background in M windows analysis of the background, the sixth input of the shaper (36) signals of the primary binary television image for a background change vector, a second input of a block (40) for determining the reliability coefficients of the signals of the secondary binary television images, a background change detector, a fourth input of the first node (42) for scaling and shifting the signals of the television image, a second input of the block (40) for determining the coefficients of reliability of the signals of the secondary binary television images detector of changes in the background, the seventh input of the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object and the second input of the block (51) for determining the coordinates and sizes, the current and average area of the generalized binary television image of the object in the analysis window, the fourth input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the fourth input of the shaper (32) signals of the binary television image of the histogram class of the modifier in the current analysis window, the fifth input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the signal generator (36) of the primary binary television image of the background change detector and the second the input of the first node (42) scaling and shifting the signals of the television image, the sixth input of the block (10) determine the current coordinates At a generalized binary television image of an object using generalized horizontal and vertical projections of a generalized binary television image of an object, it is connected to the second input of the primary binary television image signal generator (36) of the background change detector, the fifth input of the histogram binary television image signal generator (32) of the histogram classifier in the current window analysis and the third input of the block (35) determining the object / background ratio and the minimum mean square the values of the signals of the television image of the background in the M windows of analysis of the background, the seventh input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the inputs of the first buffer random access memory (30) and the second buffer random access memory (41), the first input of the node (44) for generating signals of the differential television is shown I am a moving object selector, the fifth input of the primary binary television image signal generator (36) of the background change detector and the first input of the first video data switch (31), the output of the first random access memory (30) is connected to the second input of the first video data switch (31), the first and the second outputs of which are connected to the third and sixth inputs of the shaper (32) of the binary television image signals of the histogram classifier in the current analysis window, respectively, and The fourth input of the block (35) for determining the object / background ratio and the minimum rms value of the background television image signals in M background analysis windows, the first output of which is the fifth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections generalized binary television image of the object, the first output of the shaper (32) of the signals of the binary television image of the histogram classification An ora in the current analysis window is connected to the first output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object and with the first input of the shaper (33) of horizontal and vertical projections of the signals of the binary television histogram image classifier, the output of which is connected to the first input of the shaper (50) of generalized horizontal and vertical projections signals of the generalized binary television image of the object and with the first input of the block (34) for determining the reliability coefficients of the signals of binary television images of the histogram classifier, the first output of which is connected to the second input of the generator (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the output of the shaper ( 36) the signals of the primary binary television image of the background change detector is connected to the input of the spatial of the low-pass filter (37) of the signals of the primary binary television image of the background change detector, the output of which is connected to the input of the shaper (38) of the signals of the secondary binary television image of the background change detector, the output of which is connected to the input of the shaper (39) of horizontal and vertical projections of the secondary binary signals television image of the background change detector, the output of which is connected to the third input of the generator (50) of the generalized horizontal and vertical signal projections catch a generalized binary television image of the object and with the first input of the block (40) for determining the reliability coefficients of the signals of the secondary binary television images of the background change detector, the first output of which is connected to the fourth input of the generator (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object the second buffer random access memory (41) is connected to the third input of the first node (42) scaling and shifting the signal in a television image, the output of which is connected to the input of the third buffer random access memory (43), the output of which is connected to the second input of the differential television image signal generation unit (44) of the moving objects selector, the output of which is connected to the input of the primary binary television signal generator (45) image of the moving objects selector, the output of which is connected to the input of the spatial filter (46) of the low frequencies of the signals of the primary binary television image the image of the selector of moving objects, the output of which is connected to the input of the shaper (47) of the signals of the secondary binary television image of the selector of moving objects, the output of which is connected to the input of the shaper (48) of horizontal and vertical projections of the signals of the secondary binary television image of the selector of moving objects, the output of which is connected to the first the input of the block (49) for determining the reliability coefficients of the signals of the secondary binary television images of the selector of moving objects and the sixth input of the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the first output of the block (49) for determining the reliability coefficients of the signals of the secondary binary television images of the selector of moving objects is connected to the fifth input of the shaper (50) of the generalized horizontal and vertical projections of the signals of a binary television image of an object, the first output of which is connected to the first input of the coordinate determination unit (51) and sizes, current and average area of the generalized binary television image of the object in the analysis window and with the fourth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the first and second outputs of the block ( 51) determining the coordinates and sizes, the current and average area of the generalized binary television image of the object in the analysis window are I by the second and third outputs of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, respectively, the third output of the block (51) for determining the coordinates and sizes, the current and average area of the generalized binary television image object in the analysis window is connected to the input of the block (52) analysis of the conditions for the breakdown of the automatic determination of the coordinates of the television image of the object, the first output of which is the seventh output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the third output of the block (35) for determining the object / background ratio and the minimum rms value of the signals a television image of the background in the M windows of analysis of the background is connected to the sixth output of the block (10) for determining the current coordinates of the generalized binary t the television image of the object using generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth output of which is connected to the second outputs of the binary television image signal generator (32) of the histogram classifier in the current analysis window, block (35) for determining the object / background ratio and the minimum mean square the values of the signals of the television image of the background in the M windows of the analysis of the background, block (34) determining the reliability coefficients of the signal catching binary, television images of a histogram classifier, block (40) for determining the reliability coefficients of signals of secondary binary television images of the background change detector, block (49) for determining the coefficients of reliability of signals of secondary binary television images of the selector of moving objects, block (52) for analyzing the conditions for the breakdown of automatic determination of coordinates television image of the object and shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized a binary television image of the object and the fourth output of the block (51) for determining the coordinates and sizes, the current and average area of the generalized binary television image of the object in the analysis window.

 The technical result is also achieved by the fact that the generator (32) of the binary television image signals of the histogram classifier in the current analysis window contains the generator (67) of the normalized histograms of the distribution of the brightness of the signals of the television images of the object and background and the node (68) of the formation of the binary television image of the histogram classifier, the first , the second, third, fourth and fifth inputs of the shaper (32) of signals of a binary television image of a histogram classifier in t the current analysis window is connected to the first, second, third, fourth and fifth inputs of the shaper (67) of normalized histograms of the distribution of the brightness of the television image signals of the object and background, respectively, the sixth input is connected to the second input of the node (68) for generating the binary television image of the histogram classifier, the seventh input the shaper (32) of the signals of the binary television image of the histogram classifier in the current analysis window is connected to the sixth input of the shaper (67) of the normalized histograms amm of the distribution of the brightness of the signals of the television images of the object and background and the third input of the node (68) for generating the binary television image of the histogram classifier, the first output of the shaper (67) of the normalized histograms of the distribution of the brightness of the signals of the television images of the object and background is connected to the first input of the node (68) for generating the binary television image of the histogram classifier, the output of which is the first output of the shaper (32) of signals of the binary television image the grammatical classifier in the current analysis window, the second output of the histogram distribution of the brightness of the signal brightness of the television image of the object and background is connected to the second output of the binary television image of the histogram classifier (32) in the current analysis window.

 The technical result is also achieved by the fact that the imager (36) of the signals of the primary binary television image of the detector for changing the background contains the imager (70) of the signals of the television image of the reference background, the fifth (69) and sixth (72) buffer memory, the third node (71) scaling and shifting the signals of the television image, the node (73) for generating the signals of the differential television image of the background change detector and the node (74) for generating the primary binary television image a detector for changing the background, the first and second inputs of the shaper (36) of the signals of the primary binary television image of the detector for changing the background are connected with the first and second inputs of the shaper (70) of the signals of the television image of the background and the third node (71) of scaling and shifting the signals of the television image, respectively , the third and fourth inputs are connected to the third and fourth inputs of the shaper (70) of the television image signals of the reference background, respectively, and the fifth input is connected the first input of the differential TV image signal generation unit (73) for generating a background change detector and the input of the fifth buffer random access memory (69), the output of which is connected to the fifth input of the television image signal shaper (70) of the reference background, and the sixth input of the primary signal shaper (36) the binary television image of the background change detector is connected to the third input of the node (73) for generating signals of the differential television image of the background change detector and the seventh input the house of the shaper (70) of the television image signals of the reference background, the output of which is connected to the third input of the third node (71) of scaling and shifting the signals of the television image, the output of which is connected to the input of the sixth buffer random access memory (72), the output of which is connected to the sixth input of the shaper (70) signals of a television image of a reference background and a second input of a node (73) for generating signals of a differential television image of a background change detector, the output of which is connected to the input node (74) of the formation of the primary binary television image of the background change detector, the output of which is connected to the output of the former (36) of the signals of the primary binary television image of the background change detector.

 The technical result is also achieved by the fact that the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images contains a second switch (53) of video data, a non-linear high-pass filter (54), an analyzer (55) of conditions updating the static reference television image of an object, a node (61) for generating and analyzing a type of sequences of minimum values of signals of a measure of dissimilarity of television images the fraction of rows and along the columns of the two-dimensional region for searching for displacements of the television image of the object, the third (56) and fourth (59) buffer random access memory, the imager (57) of the signals of the static and dynamic reference television images of the object, the imager (60) of signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for displacements of the television image of the object, the second node (58) for scaling and shifting the signals of the television image, the switch (62) of the data the minimum values of the signals of the measure of dissimilarity of television images, the approximator (63) of the sequence of the minimum values of the signals of the measure of dissimilarity of television images of a fourth degree polynomial, the node (64) for determining the coordinates of the television image of the object in the analysis window by the minimum position of the approximating polynomial of the fourth degree of the data of the sequence of the minimum values of measure signals dissimilarities of television images, the node (65) determining the coordinates of the television image of the object in the analysis window on the shift of the boundary of the region of rapid growth of signal values of the measure of dissimilarity of the television images relative to the center of the analysis window, the node (66) for determining the coordinates of the television image of the object in the analysis window according to the coordinates of the center of the analysis window, the first input of the block (11) for determining the coordinates of the television image object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images is connected to the first inputs of the analyzer (55) update conditions with of the static television reference image of the object and the signal shaper (60) of the measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for displacements of the television image of the object, the second input of the block (11) of determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of television images connected to the first input of the second video data switch (53), the third input of the television coordinate determination unit (11) from image of the object relative to the center of the current analysis window based on the formation of signals of a measure of dissimilarity of television images is connected to the second input of the second switch (53) of video data and to the first input of a nonlinear high-pass filter (54), the output of which is connected to the third input of the second switch (53) of video data, the fourth input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images connected n with the third input of the signal shaper (60) of the measure of dissimilarity between the signals of the television images in the two-dimensional region of the search for the displacement of the television image of the object and with the second inputs of the shaper (57) of the signals of the static and dynamic reference television images of the object and the analyzer (55) of the conditions for updating the static reference television image object, the fifth input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation the signals of the measure of dissimilarity of television images is connected to the fourth input of the imager (60) of the signals of the measure of dissimilarity between the signals of television images in the two-dimensional region for searching for displacements of the television image of the object, the sixth input of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of measure signals dissimilarities of television images is connected to the third input of the shaper (57) of the signals of the static and dynamic reference television image of the object, the seventh input of the block (11) of determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the fourth input of the shaper (57) of signals of static and dynamic reference television images of the object and to the fifth input of the shaper ( 60) signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region of the search for the displacements of the television image of the object, the eighth input d block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the second input of the nonlinear high-pass filter (54), with the fifth input of the analyzer (55) of the update conditions of the static reference television image of the object, the fourth input of the switch (62) data sequence of the minimum values of the signals of the measure of dissimilarity of television images and the seventh input of the shaper (60) signals differences between the signals of the television images in the two-dimensional region for searching for the displacements of the television image of the object, the output of the second switch (53) of the video data is connected to the second input of the shaper (60) of signals for the dissimilarity between the signals of the television images in the two-dimensional region for searching for the displacements of the television image of the object and the input of the third buffer operational a storage device (56), the output of which is connected to the first input of the shaper (57) of the signals of the static and dynamic reference television isionic images of the object, the first output of the analyzer (55) for updating the static reference television image of the object is connected to the fifth input of the shaper (57) of signals of static and dynamic reference television images of the object, the output of which is connected to the input of the second node (58) for scaling and shifting the signals of the television image the output of which is connected to the input of the fourth buffer random access memory (59), the output of which is connected to the sixth input of the driver (60) signal measures of dissimilarity between the signals of television images in the two-dimensional region for searching for displacements of the television image of the object, the output of which is connected to the input of the node (61) for generating and analyzing the type of sequences of the minimum values of the signals of the measure of dissimilarity of television images along the rows and along the columns of the two-dimensional region for searching for displacements of the television image of the object, the first the output of which is connected to the second output of the block (11) for determining the coordinates of the television image of the object relative to the center of the current its analysis window based on the formation of signals of the measure of dissimilarity of television images, the third input of the analyzer (55) of the conditions for updating the static reference television image of the object and the input of the switch (62) of the sequence of minimum values of the signals of the measure of dissimilarity of television images, the first output of which is connected to the input of the approximator (63 ) a sequence of minimum values of signals of a measure of dissimilarity of television images by a polynomial of the fourth degree, the first output of which is connected n with the input of the node (64) for determining the coordinates of the television image of the object in the analysis window according to the position of the minimum of the approximating polynomial of the fourth degree of the data sequence of the minimum values of the signals of the dissimilarity of television images, the second and third outputs of the switch (62) of the data of the sequence of the minimum values of the signals of the dissimilarity of the television images are connected with the inputs of the node (65) for determining the coordinates of the television image of the object in the analysis window by the offset of the boundary of the fast the growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window and the node (66) for determining the coordinates of the television image of the object in the analysis window according to the coordinates of the center of the analysis window, the first outputs of the node (64) for determining the coordinates of the television image of the object in the analysis window according to the minimum position of the approximating fourth polynomial of the fourth the degree of data of the sequence of minimum values of signals of the measure of dissimilarity of television images, the node (65) determining the coordinates of the television image The object of the object in the analysis window on the shift of the boundary of the region of rapid growth of signal values of the measure of dissimilarity of the television images relative to the center of the analysis window and the node (66) for determining the coordinates of the television image of the object in the analysis window according to the coordinates of the center of the analysis window are connected to the first output of the television coordinate determination unit (11) images of the object relative to the center of the current analysis window based on the formation of signals of a measure of dissimilarity of television images and with the fourth input of the analyzer (55) the conditions for updates the static reference television image of the object, the third output of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the second outputs of the node (61) for the formation and analysis of the type of sequences of the minimum values of the signals of the measure of the dissimilarity of television images along the rows and along the columns of the two-dimensional region of the search for the displacements of the television image of the object, the analyzer and (55) conditions for updating the static reference television image of the object, the approximator (63) of the sequence of the minimum values of the signals of the measure of dissimilarity of the television images of the fourth degree polynomial, the node (64) of determining the coordinates of the television image of the object in the analysis window according to the minimum position of the approximating polynomial of the fourth degree of the sequence data of the minimum the values of the signals of the measure of dissimilarity of television images, node (65) determining the coordinates of the television image CTA in the analysis window from the displacement of the boundary of rapid growth dissimilarity measure signal values of image pictures relative to the window center and analysis unit (66) determining the coordinates of the television picture of the object in the analysis window center coordinates of the analysis window.

 The technical result in terms of providing coordinates and stable image retention of the object in the center of the field of view of the video camera of the monitoring system or in the center of the tracking window when receiving both contrast and low-contrast images of stationary and moving objects and the terrain from the camera, when the object moves across the terrain, the average brightness of which coincides with the average brightness of the object, when changing the visible image of the object, due to its rotation during movement, change in illumination, obs by erecting the object in conditions of intermittent optical communication arising due to the obstruction of the object by various obstacles, such as folds of the terrain, buildings, vegetation, dust, smoke, splashes, etc., with a controlled (moving) and uncontrolled (fixed) field of view cameras of the surveillance system, in the presence of movement of the carrier of the surveillance system, causing wideband uncontrolled movements of the field of view of the camera of the surveillance system ("jitter" of the field of view of the camera of the surveillance system, which can lead to changing the relative position of the elements of the field of the television image not as a whole), with a greater dynamics of the movement of the object is achieved by the fact that in the method of determining the coordinates of the objects observed in the sequence of television images, which consists in receiving and storing signals of the current field of the television image, before receiving signals of the current n -th field of the television image, where n = 3,4,5, ..., determine the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals (n-1) - o and (n-2) -th fields of the television image, due to the action on the video camera of the monitoring system of the motion control signals of its field of view, determine the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system from the data of the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception signals of the (n-1) -th and (n-2) -th fields of a television image, simultaneously with the reception of signals of the current field of a television image, signals of an uncontrolled movement are received and stored of the field of view of the video camera of the surveillance system and use them to generate television image signals of the current frame, generate television image signals of the current nth frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between signal reception ( n-1) -th and (n-2) -th fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current The television image field, which compensates for uncontrolled movements and the roll of the field of view of the video camera of the monitoring system, forms the television image signals from the signals of the current television image frame in the current analysis window and scales them using the position and size signals of the current television image analysis window obtained by processing the previous (n-1) -th frame of the television image, and the initial values of the signals of the position and size of the current window a The analysis of the television image is formed using the initial values of the position and size of the television image of the object obtained at the start of tracking from the monitoring system, the signals of the scaled television image are stored in the current analysis window, the signals of the differential television image of the moving object selector are generated by subtraction from the signals of the scaled television image in the current window of signal analysis of a scaled television image previously remembered data and reduced to the current scale of the television image in the analysis window and offset by the displacement of the axis of the field of view of the video camera of the surveillance system, the signals of the primary binary television image of the selector of moving objects are generated from the signals of the differential television image of the selector of moving objects, the signals of the secondary binary television image of the selector of moving objects are generated from the signals of the primary binary television image of the selector of moving objects, having passed low-pass filtering, simultaneously with storing signals of a scaled television image in the current analysis window, generating signals of a differential television image of a selector of moving objects and generating signals of primary and secondary binary television images of a selector of moving objects from signals of a scaled television image in the current analysis window and taking into account the signals of a controlled moving the axis of the field of view of the video camera of the surveillance system the signals of the primary binary television image of the detector of background changes and the signals of the binary television image of the histogram classifier in the current analysis window, from the signals of the primary binary television image of the detector of background changes that have passed low-pass filtering, the signals of the secondary binary television image of the detector of background changes are formed, horizontal and vertical projections are formed signals of secondary binary television images of the moving object selector of the background and the detector of background changes and horizontal and vertical projections of the signals of the binary television image of the histogram classifier, determine the reliability coefficients of the signals of the secondary binary television images of the detector of the background change and the selector of moving objects, as well as the signals of the binary television image of the histogram classifier, form a generalized horizontal and vertical projection of the signals of the generalized binary television image of an object from horizontal and vertical projections of the signals of the secondary binary television images of the selector of moving objects and the detector of changes in the background, as well as horizontal and vertical projections of the signals of the binary television image of the histogram classifier based on their joint processing using the reliability coefficients of the signals of the binary television images of the histogram classifier, the signals of the secondary binary television images a detector of changes in the background and a selector of moving objects, divide the horizontal and vertical boundaries, as well as the dimensions of the television image of the object according to the cutoff levels left and right, above and below the specified percentage of the area of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, determine the current and average area of the generalized binary television image of the object located inside the formed boundaries of the television image of the object, determine the current coordinates of the generalized binary TV the ion image of the object, using the generalized horizontal and vertical projections of the generalized binary television image of the object, determine the current speed of the generalized binary television image of the object in the inertial coordinate system, determine the reliability coefficient of the current speed of the generalized binary television image of the object in the inertial coordinate system, simultaneously with the formation of signals primary and secondary binary television depicted the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images as a result of a nonlinear high-frequency radio frequency selector of moving objects and a detector of changes in background, binary television image of a histogram classifier, and also generalized horizontal and vertical projections of signals of a generalized binary television image of an object filtering signals of a scaled television image in the current analysis window, provided that the averaged area of the generalized binary television image of the object exceeds a threshold value, storing the received signals, generating signals of a static reference television image of an object or signals of a static and dynamic reference television image of an object, bringing signals of a static reference television image of an object or signals static and dynamic reference television images of the object to the current scale, generation and storage of signals of the measure of dissimilarity between the signals of the television images after non-linear high-frequency filtering of the signals of the scaled television image in the current analysis window and the signals of the static or static and dynamic reference television images of the object in the two-dimensional region for finding the displacements of the television image of the object, determining the minimum values of the measure signals dissimilarities of television images along rows and along two-dimensional columns of the th search area for the displacements of the television image of the object, the formation of the sequences of the minimum values of the signals of the measure of dissimilarity of the television images along the rows and the columns of the two-dimensional area of the search for the displacements of the television images of the object, the determination of the corresponding coordinate of the television image of the object in the analysis window depending on the type of the sequence of the minimum values of the signals of the measure of the dissimilarity of the television images for this coordinate, namely through analytical appr approximation of the sequence of minimum values of signals of dissimilarity measure of television images by a fourth degree polynomial and determination of the coordinate of a television image of an object as the position of a minimum approximating polynomial provided that the sequence of minimum values of signals of measure of dissimilarity of television images is assigned to the type of sequences with two boundaries of areas of rapid growth of values of signals of measure of dissimilarity of television images near provisions of its minimum, or by means of determining the displacement of the boundary of the region of rapid growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window and the formation of the object coordinate in the analysis window as a value proportional to the displacement of the boundary of the region of the fast growth of signal values of measure of dissimilarity of television images, provided that the sequence of the minimum signal values of the measure of dissimilarity of television images to the type of sequences with a flat neighborhood of the minimum position and the presence of the bottom of the region of rapid increase in the values of signals of the measure of dissimilarity of television images, or by generating the object coordinate equal to the coordinate of the center of the analysis window, subject to assigning the sequence of the minimum values of signals of the measure of dissimilarity of television images to the type of sequences with a flat neighborhood of the position of the minimum values of the signals of the measure of dissimilarity of television images in the entire region search for the displacements of the television image of the object, determine the current speed of movement of the body the image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, determine the reliability coefficient of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, form a complex Estimation of the current speed of the television image of the object inertia the coordinate system from the data of estimating the current speed of moving the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images and evaluating the current speed of moving the object from the generalized projections of the signals of the generalized binary television image of the object, taking into account the reliability coefficients of the speed components of the television image of the object and a priori object maneuvering speed limits averaged integrated assessment signals the current velocity of the television image of the object in the inertial coordinate system and remember them, determine the coordinates of the television image of the object in the field of view of the video camera of the monitoring system by integrating the difference of the complex estimate of the current speed of the television image of the object in the inertial coordinate system and the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system in inertial coordinate system, and the initial coordinates of the television image the object in the field of view of the video camera, the monitoring system and the start signal of the tracking are received from the monitoring system, simultaneously with the formation of signals of primary and secondary binary television images of the moving object selector and background change detector, binary television image of the histogram classifier, generalized horizontal and vertical projections of the signals of the generalized binary television image the object is formed by M background analysis windows, with M = 4,5,6, ..., along the perimeter of the analysis window and determine the projection of the signals of the binary television image of the histogram classifier in the M windows of analysis of the background, determine the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows of analysis of the background from the received projections of the signals of the binary television image of the histogram of the classifier in the M windows of analysis of the background, generate control signals for the movement of the axis of the field of view cameras of the surveillance system using the coordinates of the television image of the object in the field of view of the camera with observation systems obtained as a result of processing a television image in the current analysis window or using extrapolated coordinates and the moving speed of the television image of the object, depending on the results of the analysis of the current and average area of the generalized binary television image of the object, the current and average speed of moving the television image of the object, area and coordinates the boundaries of binary television images of a histogram classifier in M background analysis windows, and the extrapolated speed of moving the television image of the object is formed on the basis of the analysis of the stored values of the averaged complex estimate of the current speed of moving the television image of the object, the position and size signals of the analysis window of the television image are generated for the next frame using the coordinates of the television image of the object in the field of view of the video camera of the observation system and the dimensions of the television images of the object obtained as a result of processing the current ka pa, wherein initial values of position signals and TV image size of the object and tracking start signal is obtained from the system observation.

где X_УПР[i] , Y_УПР[i] - сигналы управления перемещением поля оси зрения видеокамеры по горизонтали и вертикали, сформированные в результате обработки i-го поля телевизионного изображения,
hx[i] - импульсная характеристика горизонтального привода оси поля зрения видеокамеры системы наблюдения, представляющая собой отклик приращения угловых координат оси поля зрения видеокамеры системы наблюдения за время между приемом i-го и (i-1)-го полей телевизионного изображения на воздействие сигнала управления по горизонтали, постоянного на интервале получения первого поля телевизионного изображения и равного нулю в другие моменты времени,
hy[i] - импульсная характеристика вертикального привода оси поля зрения видеокамеры системы наблюдения, представляющая собой отклик приращения угловых координат оси поля зрения видеокамеры системы наблюдения за время между приемом i-го и (i-1)-го полей телевизионного изображения на воздействие сигнала управления по вертикали, постоянного на интервале получения первого поля телевизионного изображения и равного нулю в другие моменты времени,
К - длина импульсной характеристики, представляющая собой число полей телевизионного изображения, по истечении которых модуль импульсной характеристики не превышает заданного уровня.The technical result is also achieved by the fact that before receiving the signals of the current n-th field of the television image, where n = 3, 4, 5, ..., the controlled movement dx [n] and dy [n] of the axis of the field of view of the video camera of the surveillance system is determined the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image horizontally and vertically, respectively, by calculating the convolution of control signals X _UPR [i], Y _UPR [i] by moving the field of view of the camera of the surveillance system with impulse characteristics hx [i] and hy [i] of its drives

where X _UPR [i], Y _UPR [i] - control signals for moving the field of view of the camera axis horizontally and vertically, formed as a result of processing the i-th field of the television image,
hx [i] is the impulse response of the horizontal axis of the field of view of the video camera of the monitoring system, which is the response of the increment of the angular coordinates of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the ith and (i-1) th fields of the television image on the influence of the control signal horizontally constant in the interval of obtaining the first field of the television image and equal to zero at other points in time,
hy [i] is the impulse response of the vertical axis of the field of view of the video camera of the surveillance system, which is the response of the increment of the angular coordinates of the axis of the field of view of the video camera of the surveillance system between the reception of the i-th and (i-1) th fields of the television image on the influence of the control signal vertically, constant in the interval of obtaining the first field of the television image and equal to zero at other points in time,
K is the length of the impulse response, which is the number of fields of the television image, after which the impulse response module does not exceed a predetermined level.

jy - номер строки в кадре, jy=1,...,Mк;
Nк - число элементов телевизионного изображения в строке;
Мк - число строк в кадре телевизионного изображения;
npk - индекс текущего полукадра:
npk = 1 - в нечетных полукадрах,
npk = 0 - в четных полукадрах,

Технический результат достигается также тем, что сигналы L_{1БИН ДИФ}(iх, jу) первичного
бинарного телевизионного изображения детектора изменений фона формируют приведением сигналов L_n-1эф(ix, jy) телевизионного изображения эталонного фона, полученных в предыдущем n-1 кадре, к текущему масштабу, образованием сигналов L_{р ДИФ}(iх, jу) разностного телевизионного изображения детектора изменений фона вычитанием из сигналов L_{n OA}(ix,jy) масштабированного изображения в текущем окне анализа сигналов L_n-1эф(ix,jy) телевизионного изображения эталонного фона предыдущего кадра со сдвигом, учитывающим перемещение Vx, Vy центра окна анализа в инерциальной системе координат за последний кадр:
L_{p ДИФ}(ix,jy) = L_{n ОА}(ix,iy) - L_{n-1 эф}(ix+Vx,jy+Vy),
определением порога бинаризации ПОРОГ _ДИФ(ix,jy) детектора изменений фона как величины, пропорциональной локальному параметру рассеяния значений сигналов разностного телевизионного изображения L_{р ДИФ}(iх,jу) детектора изменений фона в окрестности точки с координатами ix,jy, присвоением значений первичному бинарному телевизионному изображению L_{1 БИН ДИФ}(ix,jy) детектора изменений фона
L_{1 БИН ДИФ}(ix,jy)=1,
если |L_{p ДИФ}(ix,jy)|≥ПОРОГ_ДИФ(ix,jy),
или
L_{1 БИН ДИФ}(ix,jy)=0,
если |L_{p ДИФ}(ix,jy)|<ПОРОГ_ДИФ(ix,jy),
где ix,jy - координаты сигналов масштабированного телевизионного изображения относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
причем сигналы L_{n эф}(ix,jy) телевизионного изображения эталонного фона формируют разделением сигналов масштабированного телевизионного изображения в текущем окне анализа на сигналы телевизионного изображения трех типов:
сигналы телевизионного изображения в окне объекта - ОО_ДИФ,
сигналы телевизионного изображения в окне фона - ОФ_ДИФ,
сигналы телевизионного изображения в окне - "Новый фон" - НФ,
где в качестве сигналов телевизионного изображения в окне объекта определяют сигналы телевизионного изображения в прямоугольнике, находящемся в центре текущего окна анализа и включающем в себя преимущественно элементы изображения объекта, в качестве сигналов телевизионного изображения окна "Новый фон" определяют элементы изображения на внешних границах текущего окна анализа, на которых за счет движения объекта и перемещения оси поля зрения видеокамеры системы наблюдения появляются новые элементы телевизионного изображения фона, в качестве сигналов телевизионного изображения окна фона определяют все оставшиеся элементы телевизионного изображения окна анализа запоминанием в окне "Новый фон" сигналов L_{n OA}(ix,jy) из текущего окна анализа масштабированного телевизионного изображения текущего n-го кадра:
L_{n эф}(ix,jy) = L_{n OA}(ix,jy), ix,jy∈НФ,
где L_{n OA}(ix,jy) - значения сигналов яркости элемента масштабированного телевизионного изображения в окне анализа с координатами ix,jy, усреднением в окне фона сигналов масштабированного телевизионного изображения L_{n OA}(ix, jy) из текущего окна анализа с постоянной Wоф и с учетом сдвига окна анализа в инерциальной системе координат за последний кадр:
L_{n эф}(ix, jy) = (l-Wоф)*L_{n-1 эф}(iх+Vx, jy+Vy)+Wоф*L_{n OA}(ix, jy) ix, jy∈OФ_ДИФ,
где Vx, Vy - перемещение центра окна анализа за последний кадр по горизонтали и вертикали, соответственно, в инерциальной системе координат, перезаписью в окне объекта сигналов телевизионного изображения эталонного фона предыдущего кадра со сдвигом, учитывающим перемещение центра окна анализа за последний кадр:
L_{n эф}(ix,jy) = L_{n-1 эф}(iх+Vх, jy+Vy) ix,jy∈ОО_ДИФ.The technical result is also achieved by the fact that the formation of the signals L _KADR (ix, jy) of the television image of the current frame during interlacing from the signals L _P-FRAME (i, p, nрk) of the current half-frame of the television image and the television image L _-1KAAD (ix, jy ) of the previous frame is performed prediction signal L _KADR (ix, jy) of the current television picture frame by the television image displacement L _-1KADR (ix, jy) of the previous frame by the amount of managed displacement dx, dy axis of the field of view of observation system camcorder Ia horizontally and vertically, respectively, during the time between receipt of TV image half-frames
L _CADR (ix, jy) = L _-1KAADR (ix + dx, jy + dy)
and replacing the points of the television image of the current frame with the points of the television image of the current half-frame with compensation for the current uncontrolled displacements r _x (p), r _y (p) and roll φ (p) of the field of view of the video camera of the surveillance system
L _FRAME [ix (i, p, npk), jy (i, p, npk)] = L _P-FRAME (i, p, npk),
where i is the number of the element in the line of the television image of the current half-frame, i = 1, ..., Nк;
ix - element number in the line of the television image of the current frame, ix = 1, ..., Nк;
p is the line number in the half frame,

jy - line number in the frame, jy = 1, ..., Mк;
Nк is the number of elements of a television image in a row;
Mk - the number of lines in the frame of the television image;
npk - index of the current half-frame:
npk = 1 - in odd half frames,
npk = 0 - in even half frames,

The technical result is also achieved by the fact that the signals L _{1BIN DIF} (ix, jy) of the primary
the binary television image of the background change detector is formed by converting the signals L _n-1eff (ix, jy) of the television image of the reference background obtained in the previous n-1 frame to the current scale, the formation of the signals L _{r DIF} (ix, jy) of the difference television image of the change detector the background by subtracting from the signals L _{n OA} (ix, jy) the scaled image in the current signal analysis window L _n-1eff (ix, jy) of the television image of the reference background of the previous frame with a shift that takes into account the inertia of the center of the analysis window Vx, Vy coordinate system for the last frame:
L _{p DIF} (ix, jy) = L _{n ОА} (ix, iy) - L _{n-1 eff} (ix + Vx, jy + Vy),
determination of the binarization threshold _DIF threshold (ix, jy) of the background change detector as a value proportional to the local parameter of the scattering of the signals of the difference television image L _{r DIF} (ix, jy) of the background change detector in the vicinity of the point with coordinates ix, jy, assignment of values to the primary binary television image L _{1 BIN DIF} (ix, jy) of the background change detector
L _{1 BIN DIF} (ix, jy) = 1,
if | L _{p DIF} (ix, jy) | ≥POROG _DIF (ix, jy),
or
L _{1 BIN DIF} (ix, jy) = 0,
if | L _{p DIF} (ix, jy) | < _DIF Threshold (ix, jy),
where ix, jy are the coordinates of the signals of the scaled television image relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
moreover, the signals L _{n eff} (ix, jy) of the television image of the reference background are formed by dividing the signals of the scaled television image in the current analysis window into three types of television image signals:
signals of the television image in the object window - OO _DIF ,
signals of the television image in the background window - OF _DIF ,
signals of the television image in the window - "New background" - NF,
where the signals of the television image in the window of the object determine the signals of the television image in the rectangle located in the center of the current analysis window and which includes mainly the image elements of the object, as the signals of the television image of the "New background" window, image elements are determined at the external borders of the current analysis window on which new elements of the television background image appear due to the movement of the object and the displacement of the axis of the field of view of the video camera of the surveillance system, achestve television image signals background windows define all remaining elements of the television image analysis window latched in the "New background" signals L _{n OA} (ix, jy) of the current window of analysis scaled television image of the current n-th frame:
L _{n eff} (ix, jy) = L _{n OA} (ix, jy), ix, jy∈NF,
where L _{n OA} (ix, jy) are the values of the brightness signals of the scaled television image element in the analysis window with coordinates ix, jy, averaging in the background window of the signals of the scaled television image L _{n OA} (ix, jy) from the current analysis window with the constant Wof and taking into account the shift of the analysis window in the inertial coordinate system for the last frame:
L _{n eff} (ix, jy) = (l-Woff) * L _{n-1 eff} (ix + Vx, jy + Vy) + Woff * L _{n OA} (ix, jy) ix, jy∈OF _DIF ,
where Vx, Vy is the displacement of the center of the analysis window for the last frame horizontally and vertically, respectively, in the inertial coordinate system, overwriting in the window of the signal object of the television image the reference background of the previous frame with a shift that takes into account the displacement of the center of the analysis window for the last frame:
L _{n eff} (ix, jy) = L _{n-1 eff} (ix + Vx, jy + Vy) ix, jy∈OO _DIF .

где ix, jy - координаты сигналов отфильтрованного телевизионного изображения S_fil_ДИФ(iх,jу) относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
NF, MF - параметры апертуры низкочастотного фильтра по горизонтали и вертикали, соответственно;
di, dj - внутренние переменные низкочастотного фильтра по горизонтали и вертикали, соответственно, di ∈[-NF, NF], dj ∈[-MF, MF];
h_HЧ[di, dj] - импульсная характеристика низкочастотного фильтра.The technical result is also achieved by the fact that the low-pass filtering of the signals of the primary binary television image L _{1 BIN DIF} (ix, jy) of the background change detector is carried out using two-dimensional convolution

where ix, jy are the coordinates of the signals of the filtered television image S_fil _DIF (ikh, jy) relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
di, dj are the internal variables of the low-pass filter horizontally and vertically, respectively, di ∈ [-NF, NF], dj ∈ [-MF, MF];
h _HF [di, dj] - impulse response of a low-pass filter.

где ix, jy - координаты сигналов отфильтрованного телевизионного изображения S_fil_СДО(ix,jy) относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа;
NF, MF - параметры апертуры низкочастотного фильтра по горизонтали и вертикали, соответственно;
di, dj - внутренние переменные низкочастотного фильтра по горизонтали и вертикали, соответственно, di ∈[-NF, NF], dj ∈[-MF, MF];
h_HЧ[di, dj] - импульсная характеристика низкочастотного фильтра.The technical result is also achieved by the fact that the low-pass filtering of the signals of the primary binary television image L _{1 BIN SDO} (ix, jy) of the selector of moving objects is carried out using two-dimensional convolution

where ix, jy are the coordinates of the filtered television image S_fil _DLS (ix, jy) relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
di, dj are the internal variables of the low-pass filter horizontally and vertically, respectively, di ∈ [-NF, NF], dj ∈ [-MF, MF];
h _HF [di, dj] - impulse response of a low-pass filter.

NX, NY - размеры текущего окна анализа;
Роrog_ДИФ1, Роrog_ДИФ0 - значения порогов принятия решения для единичных и нулевых элементов первичного бинарного телевизионного изображения детектора изменений фона соответственно.The technical result is also achieved by the fact that the signals L _{2BIN DIF} (ix, jy) of the secondary binary television image of the background change detector are formed from the signals of the low-pass filter S_fil _DIF (ix, jу) in accordance with the rule:
L _{2 BIN DIF} (ix, jy) = 1 if S_fil _DIF (ix, iy)> Porog _DIF 1 and L _{1 BIN DIF} (ix, jy) = 1 or S_fil (ix, jy)> Porog _DIF 0 and L _{1 BIN DIF} (ix, jy) = 0,
otherwise L _{2BIN DIF} (ix, jx) = 0,
where ix, jy are the coordinates of the signals L _{2BIN DIF} (ix, jy) of the secondary binary television image of the background change detector relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
Rorog _DIF 1, Rorog _DIF 0 - values of decision thresholds for single and zero elements of the primary binary television image of the background change detector, respectively.

NX, NY - размеры текущего окна анализа,
Porog_СДО1, Porog_СДО0 - значения порогов принятия решения для единичных и нулевых элементов первичного бинарного телевизионного изображения селектора движущихся объектов соответственно.The technical result is also achieved by the fact that the signals of the secondary binary television image L _{2 BIN SDO} (ix, jy) of the selector of moving objects are formed from signals S_fil _SDO (ix, jy) of the low-pass filter in accordance with the rule:
L _{2BIN LMS} (ix, jy) = 1,
if S_fil _LMS (ix, iy)> Porog _DLS 1 and L _{1 BIN LMS} (ix, jy) = 1
or S_fil _LMS (ix, jy)> Porog _DLS 0 and L _{1 BIN LMS} (ix, jy) = 0,
otherwise _2BIN L _SDS (ix, jx) = 0,
where ix, jy are the coordinates of the signals L _{2 BIN SDO} (ix, jy) of the secondary binary television image of the selector of moving objects relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
Porog _SDO 1, Porog _SDO 0 - values of decision thresholds for single and zero elements of the primary binary television image of the selector of moving objects, respectively.

L _{БИН ГК}(iх,jу) = 0 - в противном случае,
где ix, jy - координаты сигналов бинарного телевизионного изображения L_{БИН ГК}(ix,jy) относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
L_{n OA}(ix, jy) - сигналы масштабированного телевизионного изображения в текущем окне анализа;
W_n-1 ^H _ЦОО[L_i] - нормированная гистограмма распределения яркостей L_i сигналов масштабированного телевизионного изображения в центральном окне объекта - ЦОО, полученная в предыдущем (n-1)-м кадре;
W_n-1 ^H _ОФГК[L_i] - нормированная гистограмма распределения яркостей сигналов телевизионного изображения фона в окне фона гистограммного классификатора - ОФГК, полученная в предыдущем (n-1)-м кадре;
L_i - уровень яркости сигналов телевизионного изображения;
i - номер уровня яркости сигналов телевизионного изображения, i = 1,..., N_ур;
α_n-1 - параметр, зависящий от числа элементов телевизионного изображения в окне объекта гистограммного классификатора - ООГК, классифицированных как фон в предыдущем (n-1)-м кадре;
A(ix,jy) - штрафная функция, зависящая от координат элемента ix,jy телевизионного изображения в окне анализа, в том числе при ix,jy ∈ ЦОО A(ix,jy) = А₀;
причем после определения координат Х_O, Y_O и размеров Rxo, Ryo телевизионного изображения объекта в текущем n-м кадре в качестве центрального окна объекта - ЦОО определяют прямоугольник с размерами Rх_ЦОО=Rхо, Rу_ЦОО=Rо, находящийся внутри окна анализа, центр которого совпадает с центром Хо, Yo телевизионного изображения объекта и включающий в себя преимущественно элементы телевизионного изображения объекта, в качестве окна объекта гистограммного классификатора - ООГК определяют область, заключенную между двумя прямоугольниками с общим центром и с размерами Rх_ЦОО, Rу_ЦОО и Rx_ООГК, Ry_ООГК, где Rх_ООГК>Rх_ЦОО и Rу_ООГК > Ry_ЦОО, в качестве окна фона гистограммного классификатора - ОФГК определяют область, заключенную между двумя прямоугольниками с общим центром и с размерами Rx_ООГК, Ry_ООГК и Rх_ОФГК, Rу_ОФГК, где Rх_ОФГК > Rх_ООГК и Rу_ОФГК>Ry_ООГК, по сигналам телевизионного изображения L_{n OA}(ix, jy), считываемым из окна ОФГК в текущем n-м кадре, определяют гистограмму W_ОФГК[L_i] распределения яркостей L_i изображения фона, по сигналам телевизионного изображения, выбираемым из окна ЦОО в текущем n-м кадре, определяют гистограмму W_ЦОО[L_i] распределения яркостей L_i изображения объекта, производят сглаживание гистограмм W_ОФГК[L_i] и W_ЦОО[L_i]

где h_cг[j] - импульсная характеристика сглаживающего фильтра;
(2*ns+1) - число точек импульсной характеристики сглаживающего фильтра,
определяют текущий порог П_г усреднения сглаженной гистограммы

распределения яркостей изображения объекта в соответствии с выражением

где П_г0 и k_пг - постоянные величины;
1[х] - единичная функция, определяемая условиями:
1[х] = 1 при х≥0,
1[х]=0 при х<0;
усредняют текущий порог П_г рекурсивным фильтром первого порядка

где γ_{n пор} - постоянная фильтра усреднения порога П_г, изменяющаяся в зависимости от номера кадра от значения γ_{1 пор} = 1 в первом кадре до стационарного значения γ_пор,
n - номер текущего кадра изображения;
ограничивают значения усредненного порога сверху и снизу, усредняют сглаженную гистограмму

распределения яркостей L_i телевизионного изображения объекта рекурсивным фильтром первого порядка

где γ_{n ЦОО} - постоянная фильтра усреднения гистограммы распределения яркостей телевизионного изображения объекта, изменяющаяся в зависимости от номера кадра от значения γ_{1 ЦОО} = 1 в первом кадре до стационарного значения γ_ЦОО, причем, начиная с номера кадра, превосходящего Nку,
где Nку - номер кадра, с которого усреднение гистограммы

выполняют по условиям, Nку=16,..., 128,
усреднение гистограммы

распределения яркостей L_i изображения объекта выполняют только для тех уровней яркости L_i, для которых одновременно выполняются два условия:

формируют нормированную W_n-1 ^H _ЦОО[L_i] гистограмму распределения яркостей изображения объекта из усредненной гистограммы

распределения яркостей L_i изображения объекта

где N_yp - число уровней яркости изображения L_{n OA}(ix,jy),
формируют нормированную W_n-1 ^H _ОФГКL_i] гистограмму распределения яркостей изображения фона из сглаженной гистограммы

распределения яркостей L_i изображения фона:

Технический результат достигается также тем, что горизонтальную и вертикальную проекции G_{пp СДО}(ix), V_{пр СДО}(jy) сигналов вторичного бинарного телевизионного изображения L_{2БИН СДО}(iх, jу) селектора движущихся объектов определяют в соответствии с

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где ix, jy - координаты сигналов L_{2БИН СДО}(iх,jу) вторичного бинарного телевизионного изображения селектора движущихся объектов относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа;
Nwin и Mwin - размеры вторичного бинарного изображения селектора движущихся объектов по горизонтали и вертикали соответственно.The technical result is also achieved by the fact that the signals L _{BIN GK} (ix, jy) of the binary television image of the histogram classifier are formed in accordance with the rule:
L _{BIN GK} (ix, jy) = 1 if

L _{BIN GK} (ix, jy) = 0 - otherwise,
where ix, jy are the coordinates of the signals of the binary television image L _{BIN GK} (ix, jy) relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
L _{n OA} (ix, jy) - signals of the scaled television image in the current analysis window;
W _n-1 ^H _COO [L _i ] - normalized histogram of the distribution of brightness L _{i of the} signals of the scaled television image in the central window of the object - COO obtained in the previous (n-1) -th frame;
W _n-1 ^H _OFGK [L _i ] - normalized histogram of the distribution of the brightness of the signals of the television background image in the background window of the histogram classifier - OFGK obtained in the previous (n-1) -th frame;
L _i - the brightness level of the television image signals;
i is the number of the brightness level of the television image signals, i = 1, ..., N _ur ;
α _n-1 is a parameter that depends on the number of elements of the television image in the window of the object of the histogram classifier - OOGK, classified as background in the previous (n-1) -th frame;
A (ix, jy) is a penalty function that depends on the coordinates of the element ix, jy of the television image in the analysis window, including for ix, jy ∈ CCA A (ix, jy) = A ₀ ;
moreover, after determining the coordinates X _O , Y _O and the dimensions Rxo, Ryo of the television image of the object in the current nth frame, the rectangle with dimensions Rx _ЦО = Rх, Rу _ЦОО = Rо inside the analysis window is determined as the center window of the object — COO; which coincides with the center of Ho, Yo of the television image of the object and which includes mainly the elements of the television image of the object, as the window of the object of the histogram classifier - OOGK define the area enclosed between two rectangles with a common center and with the dimensions Rx _SSC, Ru _CSB and Rx _OOGK, Ry _OOGK where Rx _OOGK> Rx _CSB and Ru _OOGK> Ry _CSB, as a window background histogram classifier - OFGK define the area contained between the two rectangles with a common center and with dimensions Rx _OOGK , Ry _OOGK and Rx _OFGK , Rу _OFGK , where Rx _OFGK > Rx _OOGK and Rу _OFGK > Ry _OOGK , from the television image L _{n OA} (ix, jy) read from the OFGK window in the current nth frame, determine the histogram W _OFGC [L _i ] the distribution of the brightness L _{i the} image of the background, according to the signals of the television image, selected from the DSP window in the current nth frame, determine the histogram W _COO [L _i ] the distribution of brightness L _{i the} image of the object, smooth the histograms W _OFGC [L _i ] and W _COO [L _i ]

where h _cg [j] is the impulse response of the smoothing filter;
(2 * ns + 1) is the number of points of the impulse response of the smoothing filter,
determine the current threshold P _g averaging the smoothed histogram

the brightness distribution of the image of the object in accordance with the expression

where P _g0 and k _pg are constant values;
1 [x] is the unit function determined by the conditions:
1 [x] = 1 for x≥0,
1 [x] = 0 for x <0;
average the current threshold P _g with a first-order recursive filter

where γ _{n pore} is the threshold averaging filter threshold P _g , which varies depending on the frame number from the value of γ _{1 pore} = 1 in the first frame to the stationary value of γ _pore ,
n is the number of the current image frame;
limit the values of the averaged threshold above and below, average the smoothed histogram

the brightness distribution L _{i of the} television image of the object with a first order recursive filter

where γ _{n COO} is the filter averaging constant of the histogram of the brightness distribution of the television image of the object, which varies depending on the frame number from the value of γ _{1 COO} = 1 in the first frame to the stationary value of the _COO , and, starting with the frame number exceeding Nku,
where Nk is the frame number from which the histogram averaging

fulfill the conditions, Nku = 16, ..., 128,
histogram averaging

the brightness distribution L _{i the} image of the object is performed only for those brightness levels L _i for which two conditions are simultaneously satisfied:

form a normalized W _n-1 ^H _COO [L _i ] histogram of the distribution of the brightness of the image of the object from the average histogram

brightness distribution L _i image of the object

where N _yp is the number of image brightness levels L _{n OA} (ix, jy),
form a normalized W _n-1 ^H _OFGC L _i ] histogram of the distribution of brightness of the background image from the smoothed histogram

luminance distribution L _i background image:

The technical result is also achieved in that the horizontal and vertical projections G _{prosp LMS} (ix), V _{ave LMS} (jy) of the secondary binary signal television image L _{2BIN LMS} (ix, ju) selector moving objects is determined according to

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where ix, jy are the coordinates of the signals L _{2 BIN SDO} (ix, jy) of the secondary binary television image of the moving objects selector relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
Nwin and Mwin are the sizes of the secondary binary image of the moving object selector horizontally and vertically, respectively.

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где ix, jy - координаты сигналов L_{2БИН ДИФ}(ix,jy) вторичного бинарного телевизионного изображения детектора изменений фона относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа;
Nwin и Mwin - размеры вторичного бинарного телевизионного изображения детектора изменений фона по горизонтали и вертикали соответственно.The technical result is also achieved in that the horizontal and vertical _straight projection _DIF G (ix), V _{ave DIF} (ju) of the secondary binary signal television image L _{2BIN DIF} (ix, ju) detector background changes determined in accordance with

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where ix, jy are the coordinates of the signals L _{2BIN DIF} (ix, jy) of the secondary binary television image of the background change detector relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
Nwin and Mwin are the sizes of the secondary binary television image of the background and horizontal changes detector, respectively.

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где ix,jy - координаты сигналов L_{БИН ГК}((ix,jy) бинарного телевизионного изображения гистограммного классификатора относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
Nwin и Mwin - размеры бинарного изображения гистограммного классификатора по горизонтали и вертикали соответственно.The technical result is also achieved by the fact that the horizontal and vertical projections G _{pr GK} (ix), V _{pr GK} (jy) of the binary image signals L _{BIN GK} (ix, jy) of the histogram classifier is determined in accordance with

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where ix, jy are the coordinates of the signals L of the _{BIN GC} ((ix, jy) of the binary television image of the histogram classifier relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
Nwin and Mwin are the sizes of the binary image of the histogram classifier horizontally and vertically, respectively.

причем усредненные плотности

бинарных телевизионных изображений детектора изменений фона, селектора движущихся объектов и гистограммного классификатора получают в результате ограничения минимальных и максимальных значений и последующего усреднения рекурсивными фильтрами первого порядка текущих плотностей V_ДИФ(n), V_СДО(n), V_ГК(n) соответствующих бинарных телевизионных изображений,
где

S_ОДИФ(n), S_ОСДО(n), S_ОГК(n) - текущие площади бинарных телевизионных изображений детектора изменений фона, селектора движущихся объектов и гистограммного классификатора, соответственно, внутри границ телевизионного изображения объекта;
S_OO(n) - текущая площадь области внутри границ телевизионного изображения объекта;
F_{НУ_ДИФ}(n), F_{НУ_СДО}(n), F _ГК(n) - функции ввода начальных условий детектора изменений фона, селектора движущихся объектов и гистограммного классификатора;
n - номер текущего кадра.The technical result is also achieved by the fact that the reliability coefficients W _DIF , W _SDO , W _{GK are} defined as the product of the input functions of the initial conditions and the normalized average densities of binary television images of the background change detector, moving object selector and histogram classifier, respectively

moreover, the averaged densities

binary television images of the background change detector, moving object selector, and histogram classifier are obtained by limiting the minimum and maximum values and subsequent averaging of the first densities V _DIF (n), V _SDO (n), V _GK (n) of the corresponding binary television images
Where

S _ODIF (n), S _OSDO (n), S _OGK (n) - current areas of binary television images of the background change detector, moving object selector and histogram classifier, respectively, inside the boundaries of the television image of the object;
S _OO (n) is the current area of the region inside the boundaries of the television image of the object;
F _{NU_DIF} (n), F _{NU_SDO} (n), F _GK (n) - input function for the initial conditions of the background change detector, moving objects selector and histogram classifier;
n is the number of the current frame.

NX, NY - размеры текущего окна анализа;
W_ГК(n), W_ДИФ(n), W_СДО(n) - коэффициенты достоверности сигналов бинарных телевизионных изображений гистограммного классификатора, сигналов вторичных бинарных телевизионных изображений детектора изменений фона и селектора движущихся объектов соответственно;
G_{пр ГК}(iх) и V_{пр ГК}(jу), G_{пр ДИФ}(iх) и V_{пр ДИФ}(jу), G_{пp СДО}(ix) и V_{пр СДО}(jу) - горизонтальные и вертикальные проекции сигналов бинарных телевизионных изображений гистограммного классификатора, сигналов вторичных бинарных телевизионных изображений детектора изменений фона и селектора движущихся объектов соответственно;
n - номер текущего кадра.The technical result is also achieved by the fact that the generalized horizontal and vertical projections _Gpro (ix, n), _Vpro (jy, n) of the signals of the generalized binary television image of the object are formed by a weighted summation of the projections of the binary television images of the background change detector, moving object selector and histogram classifier :
G _{pr O} (ix, n) = W _DIF (n) * G _{pr DIF} (ix) + W _LMS (n) * G _{pr LSS} (ix) + W _GK (n) * G _{pr GK} (ix),
V _{pr O} (jy, n) = W _DIF (n) * V _{ave DIF} (jy) + W _LMS (n) * V _{direct LMS} (jy) + W _CC (n) * V _{ave HA} (jy),
where ix, jy are the coordinates of the television image signals relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
W _ГК (n), W _ДИФ (n), W _ССО (n) - reliability coefficients of signals of binary television images of a histogram classifier, signals of secondary binary television images of a background change detector and a moving object selector, respectively;
G _{ave HA} (ix) and V _{ave HA} (ju), G _{etc. DIF} (ix) and V _{ave DIF} (ju), G _{prosp LMS} (ix) and V _{ave LMS} (ju) - horizontal and vertical projection signals binary television pictures a histogram classifier, signals of secondary binary television images of the background change detector and the moving object selector, respectively;
n is the number of the current frame.

The technical result is also achieved in that the current X coordinate _OBIN, Y _OBIN generalized binary television image of the object is determined as a weighted sum of the coordinates of the center of gravity _bct X, Y coordinates and _{bct OMED} X, Y _OMED median area television generalized binary object image
X _OBIN (n) = W _CT (n) * X _OCT ⁺ W _MED (n) * X _OMED ,
Y _OBIN (n) = W _CT (n) * Y _OCT + W _MED (n) * Y _OMED ,
W _MED (n) = 1-W _CT (n),
where W _MED (n), W _CT (n) are the weighting coefficients of the coordinates of the median of the area and the center of gravity of the generalized binary television image of the object, respectively;
n is the number of the current frame,
moreover, the weight coefficient W _CT (n) is increased with a decrease in the average deviation of the coordinates of the television image of the object from their predicted values.

The technical result is also achieved by the fact that the current horizontal VG _{OB_BIN} and vertical VV _{OB_BIN} constitute estimates of the binary moving speed
a television image of an object in an inertial coordinate system is determined in accordance with the expressions
VG _{OB_BIN} = (dX + ΔX _OA + ΔX _{OB_OA_BIN} ) / T,
VV _{OB BIN} = (dY + ΔY _OA + ΔY _{OB_OA_ BIN} ) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OB_OA_BIN} , ΔY _{OB_OA_BIN} - change the coordinates of the binary television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively.

NX, NY - размеры текущего окна анализа;
F_НФ[L] - функция двухстороннего ограничения;
F_НФ[L] = L_ПОР при L > L_ПОР;
F_НФ[L] = L при -L_ПОР ≤ L <L_ПОР;
F_НФ[L] = -L_ПОР при L < -L_ПОР;
L_ПОР - пороговый уровень функции двухстороннего ограничения;

К_НФ, К_ФВЧ - постоянные коэффициенты;
di,dj - внутренние переменные фильтра верхних частот;
NF, MF - параметры апертуры фильтра верхних частотно горизонтали и вертикали.The technical result is also achieved by the fact that non-linear high-frequency filtering of signals of a scaled television image in the current analysis window is performed in accordance with the expression
L _nNF (ix, jy) = L _nOA (ix, jy) + K _NF * F _NF [L _nFHF (ix, jy)],
where ix, jy are the coordinates of the television image signals relative to the center of the current analysis window;

NX, NY - sizes of the current analysis window;
F _NF [L] - function of two-way restriction;
F _NF [L] = L _POR for L> L _POR ;
F _NF [L] = L at -L _POR ≤ L <L _POR ;
F _NF [L] = -L _POR for L <-L _POR ;
L _POR - threshold level of the function of two-way restriction;

K _NF , K _HPF - constant coefficients;
di, dj - internal variables of the high-pass filter;
NF, MF - filter aperture parameters of the upper frequency horizontal and vertical.

The technical result is also achieved by the fact that the signals of the dynamic reference television image of the object are formed by reading in each frame the signals from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , Y _{C ET of} which determined by the difference
X _{C ET} = X _{ABOUT PZ} - X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} - Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view of the video camera of the surveillance system;
X _{OA PZ} , Y _{OA PZ} - coordinates of the center of the analysis window in the field of view of the video camera of the surveillance system.

The technical result is also achieved by the fact that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , Y _{C ET} which is determined by the difference
X _{C ET} = X _{ABOUT PZ} - X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} - Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view of the video camera of the surveillance system;
X _{OA PZ} , Y _{OA PZ} - coordinates of the center of the analysis window in the field of view of the video camera of the surveillance system;
when the conditions for changing the static reference television image of the object formed by comparing the parameters of the signals of the measures of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static and dynamic reference television images of the object, as well as comparing the parameters of the trajectories of the television image of the object obtained when using the signals static and dynamic reference television images of the object.

The technical result is also achieved by the fact that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , U _{C ET} which is determined by the difference
X _{C ET} = X _{ABOUT PZ} - X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} - Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view of the video camera of the surveillance system;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system;
when the conditions for changing the static reference television image of the object, formed on the basis of the analysis of the parameters of the signals of dissimilarity measures of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static reference television image of the object, as well as the path parameters of the television image of the object, obtained on the basis of signal analysis dissimilarity measures.

The technical result is also achieved by the fact that the current horizontal VG _{OB_NXX} and vertical VV _{OB_NXX} components of the estimation of the speed of movement of the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, are determined in accordance with the expressions
_{OB_HCX} VG = (dX + _OA + ΔH ΔH _{OB_OA NSH)} / T,
_{OB_NSKH} VV = (dY + ΔY _{OA OB_OA NSH} + ΔY) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image, horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OB_OA_NXX} , ΔY _{OB_OA_NXX} - change the coordinates of the television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, obtained on the basis of the formation of signals of the measure of dissimilarity of television images.

обобщенного бинарного телевизионного изображения, ограничения минимальных и максимальных значений текущей плотности V_Бин(n) обобщенного бинарного телевизионного изображения, последующего усреднения ограниченной плотности обобщенного бинарного телевизионного изображения рекурсивным фильтром первого порядка, нормирования усредненной плотности

обобщенного бинарного телевизионного изображения объекта

где S_ОБИН(n) - текущая площадь обобщенного бинарного телевизионного изображения внутри границ обобщенного бинарного телевизионного изображения объекта;
S_OO(n) - текущая площадь области внутри границ обобщенного бинарного телевизионного изображения объекта;
n - номер текущего кадра;

усредненный коэффициент сходства, получаемый в результате определения текущего коэффициента сходства

ограничения его максимальных и минимальных значений и усреднения рекурсивным фильтром первого порядка;
σ_фmin(n) - минимальное значение среднеквадратического значения сигналов изображения телевизионного фона в М окнах анализа фона;
Е_НСХmin(n) - минимальное значение сигналов меры несходства телевизионного изображения в текущем окне анализа и статического эталонного телевизионного изображения объекта в двумерной области поиска смещений телевизионного изображения объекта.The technical result is also achieved by the fact that the confidence coefficient W _BIN (n) of the current velocity of the generalized binary television image of the object is obtained by determining the current density

generalized binary television image, limiting the minimum and maximum values of the current density V _Bin (n) of the generalized binary television image, then averaging the limited density of the generalized binary television image with a first order recursive filter, normalizing the average density

generalized binary television image of an object

where S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object;
S _OO (n) is the current area of the region inside the boundaries of the generalized binary television image of the object;
n is the number of the current frame;

average similarity coefficient obtained by determining the current similarity coefficient

restrictions on its maximum and minimum values and averaging by a first order recursive filter;
σ _fmin (n) - the minimum value of the rms value of the image signals of the television background in M windows analysis background;
Е _НСХmin (n) - the minimum value of the signals of the measure of dissimilarity of the television image in the current analysis window and the static reference television image of the object in the two-dimensional region for searching for the displacements of the television image of the object.

ограничения его максимальных и минимальных значений и усреднения рекурсивным фильтром первого порядка, нормирования полученного усредненного коэффициента сходства

где n - номер текущего кадра;
σ_фmin(n) - минимальное значение среднеквадратического значения сигналов телевизионного изображения фона в М окнах анализа фона;
Е_НСХmin(n) - минимальное значение сигналов меры несходства телевизионных изображений текущего окна анализа и статического эталонного телевизионного изображения объекта в двумерной области поиска смещений телевизионного изображения объекта;

усредненная плотность обобщенного бинарного телевизионного изображения объекта, получаемая в результате определения текущей плотности

обобщенного бинарного телевизионного изображения объекта, ограничения минимальных и максимальных значений текущей плотности V_БИН(n) обобщенного бинарного телевизионного изображения объекта и последующего усреднения рекурсивным фильтром первого порядка ограниченной текущей плотности V_БИН(n) обобщенного бинарного телевизионного изображения объекта;
S_ОБИН(n) - текущая площадь обобщенного бинарного телевизионного изображения внутри границ обобщенного бинарного телевизионного изображения объекта;
S_ОО(n) - текущая площадь области внутри границ обобщенного бинарного телевизионного изображения объекта.The technical result is also achieved by the fact that the confidence coefficient W _CX (n) of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, is obtained by determining the current coefficient of similarity

limiting its maximum and minimum values and averaging with a first-order recursive filter, normalizing the obtained average similarity coefficient

where n is the number of the current frame;
σ _fmin (n) is the minimum rms value of the background television image signals in M background analysis windows;
Е _НСХmin (n) - the minimum value of the signals of the measure of dissimilarity of the television images of the current analysis window and the static reference television image of the object in the two-dimensional region for searching for the displacements of the television image of the object;

averaged density of a generalized binary television image of an object obtained by determining the current density

generalized binary television image of the object, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image of the object and then averaging, with a first order recursive filter, the limited current density V _BIN (n) of the generalized binary television image of the object;
S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object;
S _ОО (n) is the current area of the region inside the boundaries of the generalized binary television image of the object.

The technical result is also achieved by the fact that a comprehensive assessment of the horizontal VG _OB (n) and vertical VV _OB (n) components of the current speed of the television image of the object in the inertial coordinate system is determined by limiting the estimates of the speed of the binary television image of the object and the speed of the television image of the object, obtained on the basis of signal generation a measure of dissimilarity of television images, the minimum and maximum values formed taking into account the marching values of a comprehensive assessment of the current speed of moving a television image of an object and forming a weighted sum of limited estimates of the speed of moving a binary television image of an object and the speed of moving a television image of an object, obtained on the basis of the formation of signals of a measure of dissimilarity of television images
VG _ON (n) = W _BIN (n) * VG _{OB_BIN} (n) + W _SC (n) * VG _{HSKH ON} (n),
VV _ON (n) = W _BIN (n) * VV _{OB_BIN} (n) + W _SC (n) * VV _{HSKH ON} (n),
where W _BIN (n), W _CX (n) are the reliability coefficients of the current velocity of the generalized binary television image of the object and the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, respectively;
VG _{OB_BIN} (n) and VV _{OB_BIN} (n) - limited horizontal and vertical components of the current speed of movement of the generalized binary television image of the object;
VG _{ABOUT NSX} (n) and VV _{ABOUT NSX} (n) - limited horizontal and vertical components of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images,
n is the number of the current frame.

площади обобщенного бинарного телевизионного изображения объекта производят проверкой выполнения условия

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта или условия,

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа,
где ksl - постоянный коэффициент, ksl <1;
ks2(n-n_ЭК) - коэффициент, уменьшающийся с ростом номера кадра n, начиная с номера кадра n_ЭК, перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам, ks2(n-n_ЭК)≤ksl.The technical result is also achieved by the fact that the analysis of the current S _OBIN (n) and averaged

the area of the generalized binary television image of the object is produced by checking the fulfillment of the condition

to switch to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the extrapolated coordinates of the television image of the object or condition,

to proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window,
where ksl is a constant coefficient, ksl <1;
ks2 (nn _EC ) is a coefficient that decreases with increasing frame number n, starting from frame number n _EC , the transition to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates, ks2 (nn _EC ) ≤ksl.

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам или выполнения условий

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа,
где VG_TОБ(n), VV_TOБ(n) - горизонтальная и вертикальная составляющие текущей скорости перемещения телевизионного изображения объекта, получаемые в результате рекурсивной фильтрации первого порядка с постоянной фильтра 0<WV1<1 составляющих VG_ОБ(n), VV_ОБ(n) комплексной оценки скорости перемещения телевизионного изображения объекта:
VG_ТОБ(n) = VG_ТОБ(n-1)+ WV1*[VG_ОБ(n)-VG_ТОБ(n-1)],
VV_ТОБ(n) = VV_ТОБ(n-1)+ WV1*[VV_ОБ(n)- VV_ТОБ(n-1)],

горизонтальная и вертикальная составляющие усредненной скорости перемещения телевизионного изображения объекта, получаемые в результате рекурсивной фильтрации первого порядка с постоянной фильтра 0<WV2<1 составляющих VG_ОБ(n), VV_ОБ(n) комплексной оценки скорости перемещения телевизионного изображения объекта:

причем WV1>WV2.The technical result is also achieved by the fact that the analysis of the current and average speed of the television image of the object is carried out by checking the conditions

to go to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates or to fulfill the conditions

to proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window,
where VG _TOB (n), VV _TOB (n) are the horizontal and vertical components of the current speed of the television image of the object, obtained as a result of first-order recursive filtering with a filter constant of 0 <WV1 <1 components of VG _OB (n), VV _OB (n ) a comprehensive assessment of the speed of movement of the television image of the object:
VG _TOB (n) = VG _TOB (n-1) + WV1 * [VG _ON (n) -VG _TOB (n-1)],
_TOB VV (n) = VV _TOB (n-1) + WV1 * [VV _ON (n) - VV _TOB (n-1)],

the horizontal and vertical components of the average speed of moving the television image of the object, obtained as a result of first-order recursive filtering with a filter constant of 0 <WV2 <1 of the components VG _OB (n), VV _OB (n) of a complex estimate of the speed of the television image of the object:

moreover, WV1> WV2.

где S_i(n) - значения текущей площади бинарных телевизионных изображений гистограммного классификатора внутри i-гo окна анализа фона, i = 1,...,М;

значения усредненной площади бинарных телевизионных изображений гистограммного классификатора внутри i-гo окна анализа фона, i = 1,...,М, получаемые на выходе рекурсивного фильтра первого порядка с постоянной фильтра 0<WS<1:

Sпорог(n) - порог обнаружения изменения площади бинарных телевизионных изображений гистограммного классификатора внутри М окон анализа фона:

где k - постоянный коэффициент;

усредненная площадь обобщенного бинарного телевизионного изображения объекта, получаемая на выходе рекурсивного фильтра первого порядка с постоянной фильтра 0<WS_ОБ<1:

n - номер текущего кадра;

- отношение объект/фон,
где

среднее значение сигналов телевизионного изображения в окне объекта;

среднее значение сигналов телевизионного изображения в окне фона;
σ_Ф - среднеквадратическое отклонение сигналов телевизионного изображения в окне фона от

F_oгр(Q) - функция ограничения минимальных и максимальных значений;
при выполнении указанных выше условий для текущей площади бинарных телевизионных изображений гистограммного классификатора в одном или нескольких окнах анализа фона с номерами ki,
где ki - номера окон анализа фона, в которых обнаружено изменение текущей площади бинарного телевизионного изображения гистограммного классификатора, ki=1,...,кN;
кN - число окон анализа фона, в которых обнаружено изменение текущей площади бинарного телевизионного изображения гистограммного классификатора, кN ≤ М;
в этих окнах анализируют координаты двух взаимно перпендикулярных границ бинарного телевизионного изображения гистограммного классификатора, расположенных со стороны окна объекта, причем в окнах анализа фона, размещенных со стороны вертикальных границ окна объекта, анализируют перемещение вертикальной границы бинарного телевизионного изображения по горизонтали и контролируют принадлежность координат горизонтальной границы вертикальным координатам окна объекта, а в окнах анализа фона, размещенных со стороны горизонтальных границ окна объекта, анализируют перемещение горизонтальной границы бинарного телевизионного изображения по вертикали и контролируют принадлежность координат вертикальной границы горизонтальным координатам окна объекта, при этом анализ перемещения границ бинарных телевизионных изображений в М окнах анализа фона проводят посредством проверки условий нахождения соответствующих границ бинарного телевизионного изображения гистограммного классификатора внутри границ внутренней O_kil и внешней O_ki2 областей ki-го окна анализа фона с формированием признаков П_ki1 и П_ki2 принадлежности границ бинарных телевизионных изображений к внутренней О_ki1 и к внешней O_ki2 областям ki-го окна анализа фона,
где П_ki1= 1, если граница бинарного телевизионного изображения гистограммного классификатора лежит внутри границ области О_ki1, или П_ki1=0 - в противном случае;
П_ki2=1, если граница бинарного телевизионного изображения гистограммного классификатора лежит внутри границ области O_ki2, или П_ki2=0 - в противном случае;
причем при обнаружении переходов соответствующих границ бинарного телевизионного изображения гистограммного классификатора от внешней области к внутренней области окон анализа фона с номерами ki=kipl,
где kipl - номера окон анализа фона, в которых обнаружены переходы границ бинарного телевизионного изображения гистограммного классификатора из внешних областей O_kip12 к внутренним областям O_kip11 окон анализа фона, kipl=1,2,..., кN,
то есть при последовательном формировании признаков П_kip12=1 при n_kip1= n_kip12, а затем П_kip1=1 при n_kip1=n_kip11, где n_kip11>n_kip12, n_kip11 - номер кадра, в котором в kip1-м окне анализа фона установлено состояние признака П_kip11= 1, n_kip12 - номер кадра, в котором в kip1-м окне анализа фона установлено состояние признака П_kip12=1, и выполнении условия принадлежности координаты второй контролируемой границы бинарного телевизионного изображения гистограммного классификатора в kip1-м окне анализа фона диапазону координат окна объекта, выполняют переход к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта, присваивают счетчику SP-переходов границ бинарного телевизионного изображения гистограммного классификатора между внутренними и внешними областями окон анализа фона значение SP=1 при первом обнаружении перехода границы бинарного телевизионного изображения гистограммного классификатора из внешней области O_kip12 к внутренней области O_kip11 или увеличивают счетчик SP на единицу при повторном обнаружении переходов границы бинарного телевизионного изображения гистограммного классификатора из внешних областей O_kip12 к внутренним областям O_kip11 в течение формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта, устанавливают состояния признаков П_kip11=0, П_kip12=0,
устанавливают состояния признаков П_j2=0 в окнах анализа фона с номерами j≠kip1, j=1,...,кN, и процесс анализа границ бинарного телевизионного изображения гистограммного классификатора начинают сначала,
при обнаружении переходов соответствующих границ бинарного телевизионного изображения гистограммного классификатора из внутренних областей O_kip21 к внешним областям O_kip22 окон анализа фона,
где кiр2 - номера окон анализа фона, в которых обнаружен переход границы бинарного телевизионного изображения гистограммного классификатора из внутренних областей O_kip22 к внешних областям O_kip21 окон анализа фона, kip2= 1,2,..., кN,
в течение формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта, то есть при последовательном формировании признаков П_kiр21=1 при n_kip2=n_kip21, а затем П_kiр22=1 при n_kiр2=n_kiр22 в одном или нескольких окнах анализа фона с номерами kip2, где n_kiр22>n_kiр21, уменьшают счетчик SP-переходов границ бинарного телевизионного изображения гистограммного классификатора между внутренними и внешними областями окон анализа фона на единицу и, если SP=0, выполняют переход к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа, устанавливают состояния признаков П_kip21=0, П_kip22=0, устанавливают состояния признаков П_j2= 0 в окнах анализа фона с номерами j≠kip2, j= i, . . . , кN, и процесс анализа границ бинарного телевизионного изображения гистограммного классификатора начинают сначала, контролируют время формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам и время нахождения признаков П_i1, П_i2, i=l,...,M, в состоянии П_i1=1, П_i2=1 и при превышении установленных интервалов времени переводят соответствующие признаки в нулевое состояние, переходят к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа.The technical result is also achieved by the fact that the analysis of the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows analysis of the background is carried out by checking the conditions for detecting changes in the area of the binary television images of the histogram classifier inside the M windows of the background analysis:

where S _i (n) are the values of the current area of binary television images of the histogram classifier inside the i-th window of the background analysis, i = 1, ..., M;

values of the average area of binary television images of the histogram classifier inside the ith window of the background analysis, i = 1, ..., M, obtained at the output of the first order recursive filter with a filter constant of 0 <WS <1:

S threshold (n) - threshold for detecting a change in the area of binary television images of a histogram classifier inside M windows of background analysis:

where k is a constant coefficient;

the average area of the generalized binary television image of the object obtained at the output of a first-order recursive filter with a filter constant of 0 <WS _OB <1:

n is the number of the current frame;

- object / background ratio,
Where

the average value of the television image signals in the object window;

the average value of the television image signals in the background window;
σ _f - the standard deviation of the signals of the television image in the background window from

F _ogr (Q) - the function of limiting the minimum and maximum values;
when the above conditions are met for the current area of binary television images of the histogram classifier in one or more background analysis windows with numbers ki,
where ki are the numbers of the background analysis windows in which a change in the current area of the binary television image of the histogram classifier was detected, ki = 1, ..., kN;
kN is the number of background analysis windows in which a change in the current area of the binary television image of the histogram classifier was detected, kN ≤ M;
in these windows, the coordinates of two mutually perpendicular boundaries of the binary television image of the histogram classifier located on the side of the object window are analyzed, and in the background analysis windows located on the side of the vertical borders of the object window, the horizontal movement of the vertical border of the binary television image is analyzed and the horizontal coordinate coordinates are controlled the vertical coordinates of the object’s window, and in the background analysis windows placed on the side of horizontal faces the object’s windows, analyze the vertical horizontal displacement of the binary television image and control the vertical border coordinates of the horizontal coordinates of the object’s window, while the analysis of the displacement of the binary television image boundaries in the M analysis windows is carried out by checking the conditions for finding the corresponding boundaries of the histogram binary television image within the boundaries internal O _ki l and external O _ki 2 areas of the ki-th background analysis window with knowledge of the signs P _ki 1 and P _ki 2 that the boundaries of the binary television images belong to the internal O _ki 1 and to the external O _ki 2 regions of the ki-th background analysis window
where П _ki 1 = 1, if the boundary of the binary television image of the histogram classifier lies inside the boundaries of the region О _ki 1, or П _ki 1 = 0 otherwise;
П _ki 2 = 1, if the boundary of the binary television image of the histogram classifier lies within the boundaries of the region O _ki 2, or П _ki 2 = 0 otherwise;
moreover, when detecting transitions of the corresponding boundaries of the binary television image of the histogram classifier from the outer region to the inner region of the background analysis windows with numbers ki = kipl,
where kipl are the numbers of the background analysis windows in which transitions of the boundaries of the binary television image of the histogram classifier from the outer regions O _kip1 2 to the inner regions O _kip1 1 of the background analysis windows, kipl = 1,2, ..., kN,
that is, in the sequential formation of the attributes P _kip1 2 = 1 for n _kip1 = n _kip1 2, and then P _kip 1 = 1 for n _kip1 = n _kip1 1, where n _kip1 1> n _kip1 2, n _kip1 1 is the frame number, in which in the kip1st window of the background analysis the status of the attribute П _kip1 1 = 1 is set, n _kip1 2 is the frame number, in which in the kip1th window of the analysis of the background the state of the sign П _kip1 2 = 1 is set, and the condition of the second coordinate being controlled is satisfied the boundaries of the binary television image of the histogram classifier in the kip1st window of the background analysis to the coordinate range of the object window, go to the formation of In order to control the movement of the axis of the field of view of the video camera of the monitoring system according to the extrapolated coordinates of the television image of the object, set the counter of SP transitions of the boundaries of the binary television image of the histogram classifier between the internal and external regions of the background analysis windows to SP = 1 upon the first detection of the transition of the boundary of the binary television image of the histogram classifier from O _kip1 outer region to the inner region 2 O _kip1 1 or increase the counter by one when SP repeated bnaruzhenii border crossings binary television image histogram classifier from external domains O _kip1 2 to the interior of O _kip1 1 during the formation of the axis movement control signals of the field of view of video camera surveillance system according to the extrapolated coordinates of the television picture of the object, sets the state attributes P _kip1 1 = 0, n _kip1 2 = 0,
establish the state of signs П _j 2 = 0 in the background analysis windows with numbers j ≠ kip1, j = 1, ..., кN, and the process of analyzing the boundaries of the binary television image of the histogram classifier starts again,
upon detecting transitions of the corresponding boundaries of the binary television image of the histogram classifier from the inner regions O _kip2 1 to the outer regions O _kip2 2 of the background analysis windows,
where kir2 are the numbers of the background analysis windows in which the transition of the boundary of the binary television image of the histogram classifier from the inner regions O _kip2 2 to the outer regions O _kip2 1 of the background analysis windows, kip2 = 1,2, ..., kN,
during the generation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by the extrapolated coordinates of the television image of the object, that is, with the sequential formation of signs P _kip2 1 = 1 for n _kip2 = n _kip2 1 and then P _kip2 2 = 1 for n _kip2 = n _kip2 2 in one or several background analysis windows with numbers kip2, where n _kip2 2> n _kip2 1, reduce the counter of SP transitions of the boundaries of the binary television image of the histogram classifier between the internal and external regions of the background analysis windows by one and, if SP = 0, weeding they take the transition to the formation of control signals for moving the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window, set the status of signs П _kip2 1 = 0, П _kip2 2 = 0 , establish the state of signs П _j 2 = 0 in the background analysis windows with numbers j ≠ kip2, j = i,. . . , kN, and the process of analyzing the boundaries of the binary television image of the histogram classifier begin first, control the time of formation of the control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates and the time of finding signs P _i 1, P _i 2, i = l, ..., M, in the state П _i 1 = 1, П _i 2 = 1 and when the set time intervals are exceeded, the corresponding signs are transferred to the zero state, they proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the observation system the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window.

The technical result in terms of providing coordinates and stable image retention of the object in the center of the field of view of the video camera of the monitoring system or in the center of the tracking window when receiving from the video camera of the monitoring system both contrast and low-contrast images of stationary and moving objects and the background of the terrain, when the object moves across sections terrain, the average brightness of which coincides with the average brightness of the object, when the visible image of the object changes due to its rotation during movement, the change illumination of the object under observation intermittent optical communications arising from the overlap of the object with various obstacles, such as the folds of the terrain, buildings, vegetation, dust, smoke, spray, and so on. P. , with a controlled (mobile) and uncontrolled (fixed) field of view of the video camera of the surveillance system, if there is movement of the carrier of the surveillance system causing wideband uncontrolled movements of the field of view of the video camera of the monitoring system ("jitter" of the field of view of the video camera of the monitoring system, which can lead to a change in relative position field elements of the television image is not as a whole), with greater dynamics of the movement of the object is achieved by the fact that in the device for determining the coordinates of objects, observing In the sequence of television images, containing a processor (95) for processing local data transmitted via a common bi-directional bus, a unit (75) for receiving and storing signals of the current field of a television image from a video camera of the monitoring system and generating signals for synchronizing the operation of the device, block (76) receiving and storing signals of uncontrolled movement and roll of the field of view of the video camera of the surveillance system, block (77) for calculating the controlled movement of the axis of the field of view of the video camera system we observation time between reception signals (n-1) th and (n-2) -th fields of the television image, where n = 3, 4, 5 ,. . . - the number of the current field of the television image, the imager (78) of the signals of the television image of the current n-th frame from the signals of the television image of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals of the (n-1) -th and (n -2) th fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movement measurements and roll of the field of view of the video camera of the monitoring system, a unit (79) for generating and scaling television image signals in the current analysis window using position and size signals of the current television image analysis window, a unit (80) for generating M background analysis windows around the perimeter of the analysis and determination window projections of signals of a binary television image of a histogram classifier in M background analysis windows, block (81) for determining the current coordinates of a generalized binary television image of an object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images, the block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object obtained on the basis of signal generation measures of dissimilarity of television images, the switch (84) codes of the source or current dimensions of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, the block (85) for determining the reliability coefficients of the current speed of the generalized binary television image of the object in the inertial coordinate system, block ( 86) determining the reliability coefficients of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of generating signals of a measure of dissimilarity of television images, a block (87) for determining the area and coordinates of the boundaries of binary television images in M background analysis windows, a block (88) for determining the current velocity of a generalized binary television image of an object in an inertial coordinate system, a unit for generating an integrated estimate the current speed of movement of the television image of the object in an inertial coordinate system, block (90) of an integrated estimate of the coordinates of the television image of the object in the field ia the video camera of the monitoring system, the analyzer (91) of the conditions for the transition to using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in M windows analysis background, block (92) averaging a comprehensive assessment of the current speed of movement of the television image the object and storing the values of the average speed, the unit (93) for determining the predicted coordinates and the speed of the television image of the object in the next frame based on the analysis of the stored values of the average complex estimate of the speed of the television image of the object, the generator (94) of signals for controlling the movement of the axis of the field of view of the video camera of the system observation and signals of the position and size of the analysis window in the next frame of the television image using the coordinates of the television about the image of the object in the field of view of the video camera of the monitoring system or the extrapolated coordinates and the speed of the television image of the object, the first and second inputs of the device connected to the first and second inputs of the block (75) receiving and storing signals of the current field of the television image from the video camera of the monitoring and signal generation system synchronize the operation of the device, respectively, the third input of the device is connected to the input of the block (76) for receiving and storing signals of uncontrolled movement and bank For viewing the video camera of the surveillance system, the fourth and fifth inputs of the device are connected to the first and second inputs of the switch (84) of the codes of the initial or current sizes of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, respectively, the sixth input of the device is connected to the second input of the block (77 ) calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image, where n = 3, 4, 5 ,. . . - the number of the current field of the television image, the first output of the block (75) receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device is connected to the first input of the imager (78) of the signals of the television image of the current n-th frame from the signals of the television images of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television from signals and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the monitoring system, the second output of the block (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring and signal generation system synchronization of the operation of the device is connected to the third input of the block (77) for calculating the controlled movement of the axis of the video field of view amers of the monitoring system for the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image, where n = 3, 4, 5 ,. . . - the number of the current field of the television image, the fifth input of the block (79) generating and scaling the signals of the television image in the current analysis window using the position and size signals of the current analysis window of the television image, the fourth input of the imager (78) of the television image signals of the current n-th frame from signals of the television image of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between reception of signals of the (n-1) -th and (n-2) -th the television image and the signals of the current field of the television image with the elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, by the second inputs of the block (80) for the formation of M windows for analysis of the background around the perimeter of the window for analysis and determination of signal projections of a binary television image of a histogram classifier in M windows for analysis of the background and block (87) for determining the area and coordinates of the gran of binary television images in the M windows of the background analysis, the third input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth input of the block (82) for determining the coordinates of the television image of the object relative to the center the current analysis window based on the formation of signals of a measure of dissimilarity of television images, the fourth input of the block (12) is determined the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, the fifth input of the switch (84) codes of the source or current dimensions of the television image of the object and the coordinates of the object in the field of view of the camera system observation, the fourth inputs of block (85) determine the reliability coefficients of the current velocity of the generalized binary about a television image of an object in an inertial coordinate system, and a block (86) for determining the reliability coefficients of the current speed of moving a television image of an object in an inertial coordinate system, obtained on the basis of generating signals of a measure of dissimilarity of television images, with the fourth input of a block (88) for determining the current movement speed of a generalized binary television image of the object in an inertial coordinate system, the fifth input of the block (89) forming a complex estimate of the current speed the movement of the television image of the object in the inertial coordinate system, the third input of the block (90) for the complex assessment of the coordinates of the television image of the object in the field of view of the video camera of the monitoring system, the sixth input of the analyzer (91) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area generalized binary television image of an object, current and average speed of a television image of an object, area the coordinates of the boundaries of the binary television images of the histogram classifier in M background analysis windows, the third inputs of the unit (92) for averaging a complex estimate of the current speed of the television image of the object and storing the values of the average speed, and the unit (93) for determining the predicted coordinates and speed of the television image of the object in the following frame based on the analysis of the stored values of the averaged integrated estimate of the speed of movement of the television image of the object, and the sixth the input of the shaper (94) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and the speed of the television image of the object, the first output block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals (n- 1) th and (n-2) th fields of the television image, where n = 3, 4, 5,. . . - number of the current field of the television image, connected to the second input of the shaper (78) of the television image signals of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals (n-1) -go and (n-2) fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, the output of the block (76) for receiving and storing signals of uncontrolled movement and the roll of the field of view of the video camera of the monitoring system is connected to the third input of the former (78) of the television image signals of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image and the signals of the current field of the television image with element-wise conversion of the s channel of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, with the fifth input of the block (81) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the first input of the block (82) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, with the first input of the block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, with the second input of the block (88) determining the current velocity of the generalized binary television image of the object in an inertial coordinate system, and with the first input of the block (90) of an integrated assessment of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the output of the shaper (78) of the television image signals of the current n-th frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the monitoring system is connected to the first input of the block (79) for generating and scaling television image signals in the current analysis window using position and size signals of the current television image analysis window, the output of the television image signal generation and scaling unit (79) in the current analysis window using the position and size signals of the current television image analysis window is connected to the seventh input of the unit (81) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of the generalized binary television image of the object and with the third input of the block (82) determining the coordinates of the television image of the object and relative to the center of the current analysis window based on the formation of signals of a measure of dissimilarity of television images, the output of the block (80) for the formation of M windows for analysis of the background around the perimeter of the window for analyzing the signal projections of the binary television image of the histogram classifier in the M windows for analysis of the background is connected to the first input of the block (87) for determining the area and coordinates of the boundaries of the binary television images in the M windows for analysis of the background, the first output of which is connected to the third input of the analyzer (91) of the transition conditions to use the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the first output of the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the block (80) for the formation of M background analysis windows along the perimeter of the analysis window and for determining the projections of the signals of the binary television image histogram classifier in M background analysis windows, the second output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the switch (84) of the codes of the initial or current sizes of the television image of the object and the coordinates of the object in the field of view of the camera of the system observations the first input of the block (85) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in the inertial coordinate system and the first input of the shaper (94) of the control signals for the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image with using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system or extrapolated coordinates and awns moving television image of the object, the third output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the block (88) for determining the current speed of the movement of the generalized binary television image of the object in the inertial coordinate system, the fourth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the second input of the block (85) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in the inertial coordinate system and with the second input of the analyzer (91), the conditions for the transition to the use of extrapolated coordinates evizionnogo image of the object based on the current and the average area of a generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the fifth output of the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the analyzer (91) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of an object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the sixth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the block (86) for determining the reliability coefficients of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the seventh output of the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the fourth input of the block (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating synchronization signals device operation the first output of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the second input of the block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the third input of the block (89) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the second output of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the first input of the block (86) for determining the reliability coefficients of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the second input of the block (89) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the first output of the switch (84) of codes of the source or current dimensions of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the first input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image facility the seventh input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the third input of the block (79) for generating and scaling the signals of the television image in the current analysis window using the position and size signals of the current analysis window television image the second output of the switch (84) of codes of the initial or current dimensions of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the second input of the block (81) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of the generalized binary television image facility the sixth input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the second input of the block (79) for generating and scaling the signals of the television image in the current analysis window using the position and size signals of the current analysis window television image the first output of the block (85) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in the inertial coordinate system is connected to the first input of the block (89) for generating a complex estimate of the current speed of the television image of the object in the inertial coordinate system, and the second output is connected to the second input of the block (86) for determining the reliability coefficients of the current speed of movement of the television image of the object in an inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the second output of which is connected to the third input of the block (85) for determining the reliability coefficients of the current velocity of the generalized binary television image of the object in an inertial coordinate system, the first output of the block (88) for determining the current moving speed of the generalized binary television image of the object in the inertial coordinate system is connected to the fourth input of the block (89) for generating a complex estimate of the current speed of moving the television image of the object in the inertial coordinate system, the first output of the unit (89) for generating a comprehensive estimate of the current speed of the television image of the object in the inertial coordinate system is connected to the fourth input of the analyzer (91) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows, the fifth input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, the first input of the unit (92) for averaging a complex estimate of the current speed of moving a television image of an object and storing the values of the average speed and the second input of a block (90) for a complex estimate of the coordinates of a television image of an object in the field of view of a video camera of a surveillance system, the first output of the unit (90) for the complex assessment of the coordinates of the television image of the object in the field of view of the video camera of the monitoring system is connected to the fourth input of the switch (84) codes of the initial or current sizes of the television image of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, the first input of the unit (93) for determining the predicted coordinates and the speed of moving the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of moving the television image of the object and the third input of the driver (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the observation system and position signals and dimensions of the analysis window in the next frame of the television image using the coordinates of the television image of the object in Beyond the view of the video camera of the surveillance system or extrapolated coordinates and the speed of the television image of the object the second output of which is connected to the third inputs of the block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of the formation of the signals of the measure of dissimilarity of the television images and the unit (87) for determining the current speed of movement of the generalized binary television image of the object in an inertial coordinate system, the sixth input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object and the fourth inputs of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of television dissimilarity images and block (79) of the formation and scaling of television image signals in the current window analysis using position and size signals of the current television image analysis window, the first output of the analyzer (91) of the transition conditions to use the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows of analysis of the background is connected to the second input of the generator (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the observation system or extrapolated coordinates and the speed of movement of the television image of the object, the second input of the block (92) averaging a comprehensive estimate of the current speed of the television image of the object and storing the values of the average speed and with the fourth input of the block (81) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the second output of the analyzer (91) of the transition conditions to use the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows of analysis of the background is connected to the second input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images, the fourth output of the analyzer (91) of the transition conditions to use the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows of analysis of the background is connected to the third input of the block (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device, the first output of the averaging unit (92) for complex estimation of the current speed of moving the television image of the object and storing the values of the average speed is connected to the fifth input of the analyzer (91) of the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in M background analysis windows and the second input of the block (93) for determining the predicted coordinates and speed of the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of the television image of the object, the second output of the unit (93) for determining the predicted coordinates and the speed of moving the television image of the object in the next frame based on the analysis of the stored values of the averaged complex estimate of the speed of moving the television image of the object is connected to the fifth input of the driver (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and signals the position and size of the analysis window in the next frame of the television image using the coordinates of the television image that in the system of surveillance cameras or extrapolated coordinates and velocity of the television image of the object, the first output of the block (93) determining the predicted coordinates and speed of the television image of the object in the next frame based on the analysis
the stored values of the averaged complex estimate of the speed of moving the television image of the object is connected to the fourth input of the shaper (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the system observation or extrapolated coordinates and speed of the television image of the object, first in the output of the driver (94) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and the speed of movement of the television image of the object is the output of the device and connected to the first input of the block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the surveillance system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image, where n = 3, 4, 5, ... is the number of the current field of the television image, the processor output (95) is computational processing local data transmitted via a common bi-directional bus is connected to the second outputs of the block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of a television image , where n = 3, 4, 5, ... is the number of the current field of the television image, block (82) for determining the ordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images, the unit (83) for determining the current speed of movement of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, block (87) determining the area and coordinates of the boundaries of binary television images in M background analysis windows a, block (89) for the formation of a complex estimate of the current speed of moving a television image of an object in an inertial coordinate system, block (90) for a complex estimate of the coordinates of a television image of an object in the field of view of a video camera of a surveillance system, block (93) for determining the predicted coordinates and speed of a TV image of an object in the next frame, based on the analysis of the stored values of the averaged complex estimate of the speed of movement of the television image of the object, the third outputs are a (85) determining the reliability coefficients of the current speed of movement of the generalized binary television image of the object in the inertial coordinate system, block (86) determining the reliability coefficients of the current speed of moving the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, the analyzer ( 91) the conditions for the transition to the use of extrapolated coordinates of the television image of the object based on the current the average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows for analysis of the background, the unit (92) for averaging the complex estimate of the current speed of the television image of the object and remembering the values of the average speed and shaper (94) of signals for controlling the movement of the axis of the field of view of the video camera of the surveillance system and signals the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and the speed of the television image of the object, the eighth output of the block (75) receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating synchronization signals of the device.

 The technical result is also achieved by the fact that the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object contains a first video data switch (31), a first buffer random access memory (30), and a second buffer random access memory (41) and a third buffer random access memory (43), a shaper (32) of binary television signals histogram classifier images in the current analysis window, primary binary television image signal shaper (36) for background change detector, secondary binary television image signal shaper (38) for background changes detector, signal shaper (45) for primary binary television image of moving objects selector, shaper (47 ) signals of the secondary binary television image of the selector of moving objects, the first node (42) scaling and shifting the signals of the body of a visual image, a node (44) for generating signals of a differential television image of a selector of moving objects, a shaper (33) of horizontal and vertical projections of signals of a binary television image of a histogram classifier, a shaper (39) of horizontal and vertical projections of signals of a secondary binary television image of a detector of background changes, a shaper ( 48) horizontal and vertical projections of the signals of the secondary binary television image of the moving selector objects, block (35) for determining the object / background ratio and the minimum rms value of the background television image signals in M background analysis windows, a spatial filter (37) of the low frequencies of the signals of the primary binary television image of the background change detector, spatial filter (46) of the low signal frequencies primary binary television image of the moving object selector, block (34) for determining the reliability coefficients of signals of binary television images of a histogram classifier a torus, a block (40) for determining the reliability coefficients of the signals of the secondary binary television images of the background change detector, a block (49) for determining the reliability coefficients of the signals of the secondary binary television images of the selector of moving objects and a generator (50) of generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, block (51) for determining coordinates and sizes, current and average area of a generalized binary television image object and in the analysis window and the block (52) for analyzing the breakdown conditions for automatically determining the coordinates of the television image of the object, the first and second inputs of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object are connected to the first and second inputs of the shaper (32) of the binary television image signals of the histogram classifier in the current analysis and block window (35) determining the object / background ratio and the minimum rms value of the background television image signals in M background analysis windows and the fourth and third inputs of the primary binary television image signal detector of the background change detector, respectively, 36, the first input of the generalized coordinates coordinates determining unit (81) a binary television image of an object using generalized horizontal and vertical projections of a generalized binary television image of an object is connected also with the first input of the first node (42) for scaling and shifting the television image signals, the third input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the first switch (31 ) of video data, the seventh input of the shaper (32) of the binary television image signals of the histogram classifier in the current analysis window, the second the shaper (33) of horizontal and vertical projections of the binary television image signals of the histogram classifier, the second input of the block (34) for determining the reliability coefficients of the signals of the binary television images of the histogram classifier, the fifth input of the block (35) for determining the object / background ratio and the minimum rms value of the television image signals background in M windows analysis of the background, the sixth input of the shaper (36) signals of the primary binary television image de a background change vector, a second input of a block (40) for determining the reliability coefficients of the signals of the secondary binary television images, a background change detector, a fourth input of the first node (42) for scaling and shifting the signals of the television image, a second input of the block (40) for determining the coefficients of reliability of the signals of the secondary binary television images detector of changes in the background, the seventh input of the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television the second image of the object and the second input of the block (51) for determining the coordinates and sizes, the current and average area of the generalized binary television image of the object in the analysis window, the fourth input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the fourth input of the shaper (32) signals of the binary television image of the histogram class of the indicator in the current analysis window, the fifth input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the signal generator (36) of the primary binary television image of the background change detector and the second the input of the first node (42) scaling and shifting the signals of the television image, the sixth input of the block (81) determine the current coordinates At a generalized binary television image of an object using generalized horizontal and vertical projections of a generalized binary television image of an object, it is connected to the second input of the primary binary television image signal generator (36) of the background change detector, the fifth input of the histogram binary television image signal generator (32) of the histogram classifier in the current window analysis and the third input of the block (35) determining the object / background ratio and the minimum root mean square the values of the signals of the television image of the background in the M windows of analysis of the background, the seventh input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the inputs of the first buffer random access memory (30) and the second buffer random access memory (41), the first input of the node (44) for generating signals of the differential television image a moving object selector, the fifth input of the primary binary television image signal generator (36) of the background change detector and the first input of the first video data switch (31), the output of the first buffer random access memory (30) is connected to the second input of the first video data switch (31), the first and the second outputs of which are connected to the third and sixth inputs of the shaper (32) of the binary television image signals of the histogram classifier in the current analysis window, respectively, and with four the first input of the block (35) for determining the object / background ratio and the minimum rms value of the background television image signals in M background analysis windows, the first output of which is the fifth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections generalized binary television image of the object, the first output of the shaper (32) of the signals of the binary television image of the histogram classifier and in the current analysis window it is connected to the first output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object and with the first input of the shaper (33) of horizontal and vertical projections of the histogram binary television image classifier, the output of which is connected to the first input of the shaper (50) of generalized horizontal and vertical projections with ignals of the generalized binary television image of the object and with the first input of the block (34) for determining the reliability coefficients of the signals of the binary television images of the histogram classifier, the first output of which is connected to the second input of the generator (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the output of the shaper ( 36) the signals of the primary binary television image of the background change detector is connected to the input of the spatial the low-pass filter (37) of the signals of the primary binary television image of the background change detector, the output of which is connected to the input of the shaper (38) of the signals of the secondary binary television image of the background change detector, the output of which is connected to the input of the shaper (39) of horizontal and vertical projections of the secondary binary television signals images of a background change detector, the output of which is connected to the third input of the shaper (50) of the generalized horizontal and vertical signal projections a generalized binary television image of the object and with the first input of the block (40) for determining the reliability coefficients of the signals of the secondary binary television images of the background change detector, the first output of which is connected to the fourth input of the generator (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the output of the second buffer random access memory (41) is connected to the third input of the first node (42) scaling and shifting signals a live image, the output of which is connected to the input of the third buffer random access memory (43), the output of which is connected to the second input of the differential television image signal generation unit (44) of the moving objects selector, the output of which is connected to the input of the primary binary television image signal generator (45) a moving object selector, the output of which is connected to the input of the spatial filter (46) of the low frequencies of the signals of the primary binary television image selector of moving objects, the output of which is connected to the input of the shaper (47) of signals of the secondary binary television image of the selector of moving objects, the output of which is connected to the input of the shaper (48) of horizontal and vertical projections of the signals of the secondary binary television image of the selector of moving objects, the output of which is connected to the first input block (49) determining the reliability coefficients of the signals of the secondary binary television images of the selector of moving objects and the sixth input the shaper house (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the first output of the block (49) for determining the reliability coefficients of the signals of the secondary binary television images of the selector of moving objects is connected to the fifth input of the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the first output of which is connected to the first input of the unit (51) for determining the coordinates and size , the current and average area of the generalized binary television image of the object in the analysis window and with the fourth output of the block (81) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the first and second outputs of the block (51) determining the coordinates and sizes, the current and average area of the generalized binary television image of the object in the analysis window are the second and the third outputs of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, respectively, the third output of the block (51) for determining the coordinates and sizes, the current and average area of the generalized binary television image of the object in the window the analysis is connected to the input of the block (52) analysis of the conditions for the failure of the automatic determination of the coordinates of the television image of the object the one whose first output is the seventh output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the third output of the block (35) for determining the object / background ratio and the minimum rms value of the television signals background images in M windows of analysis of the background is connected to the sixth output of the block (81) for determining the current coordinates of the generalized binary television of the image of the object using generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth output of which is connected to the second outputs of the generator of binary television image signals (32) of the histogram classifier in the current analysis window, block (35) for determining the object / background ratio and the minimum mean square the values of the signals of the television image of the background in the M windows of the analysis of the background, the block (34) for determining the reliability coefficients of the binary signals x television images of the histogram classifier, block (40) for determining the reliability coefficients of signals of the secondary binary television images of the background change detector, block (49) for determining the coefficients of reliability of signals of the secondary binary television images of the selector of moving objects, block (52) for analyzing the conditions for the breakdown of the automatic determination of the coordinates of the television image object and shaper (50) of generalized horizontal and vertical projections of signals of a generalized binary t a televised image of the object and the fourth output block (51) determining the coordinates and sizes, and the current average area generalized television binary object image in the analysis window.

 The technical result is also achieved by the fact that the shaper (32) of the binary television image signals of the histogram classifier in the current window of the block analysis (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object contains the shaper (67) normalized histograms of the distribution of the brightness of the signals of the television images of the object and background and the node (68) of the a binary television image of a histogram classifier, the first, second, third, fourth and fifth inputs of the generator (32) of signals of a binary television image of a histogram classifier in the current analysis window connected to the first, second, third, fourth and fifth inputs of the generator (67) of normalized histograms the distribution of the brightness of the signals of the television images of the object and the background, respectively, the sixth input is connected to the second input of the node (68) forming a binary television image g the grammatical classifier, the seventh input of the binary television image signal generator (32) of the histogram classifier in the current analysis window is connected to the sixth input of the normalized histogram of the distribution of signal brightness of the object and background television images and the third input of the binary television image of the histogram classifier (68) , the first output of the shaper (67) of normalized histograms of the distribution of the brightness of the signals of the television images of the object and is connected to the first input of the histogram classifier binary TV image generation unit (68), the output of which is the first output of the histogram classifier binary television image generator (32) signals in the current analysis window, the second output of the histogram classifier (67) normalized histogram distribution of the television image image signal brightness and the background is connected to the second output of the shaper (32) of the binary television image signals of the histogram classifier in the current m analysis window.

 The technical result is also achieved by the fact that the shaper (36) of the signals of the primary binary television image of the background change detector of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object contains the shaper (70) of television signals images of the reference background, the fifth (69) and sixth (72) buffer random access memory devices, the third node (71) m stapling and shifting the signals of the television image, the node (73) for generating the signals of the differential television image of the detector for changing the background and the node (74) for generating the primary binary television image for the detector for changing the background, the first and second inputs of the generator (36) of signals of the primary binary television image for the detector for changing the background connected to the first and second inputs of the shaper (70) of the television image signals of the reference background and the third node (71) of scaling and shifting with of the television image signals, respectively, the third and fourth inputs are connected to the third and fourth inputs of the shaper (70) of the television image signals of the reference background, respectively, and the fifth input is connected to the first input of the differential TV image signal generation unit (73) of the background change detector and the fifth buffer operational input a storage device (69), the output of which is connected to the fifth input of the former (70) of the television image signals of the reference background, and the sixth input is formed The spell (36) of the signals of the primary binary television image of the background change detector is connected to the third input of the node (73) for generating the difference television image of the background change detector and the seventh input of the signal generator of the television image 70 (reference), the output of which is connected to the third input of the third node ( 71) scaling and shifting the signals of a television image, the output of which is connected to the input of the sixth buffer random access memory (72), the output of which is connected with the sixth input of the shaper (70) of the television image signals of the reference background and the second input of the node (73) for generating the signals of the differential television image of the background change detector, the output of which is connected to the input of the node (74) of the formation of the primary binary television image of the background change detector, the output of which is connected to the output of the shaper (36) of the signals of the primary binary television image detector changes in the background.

 The technical result is also achieved by the fact that the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images contains a second switch (53) of video data, a nonlinear high-pass filter (54), an analyzer (55) of conditions updating the static reference television image of an object, a node (61) for generating and analyzing a type of sequences of minimum values of signals of a measure of dissimilarity of television images the fraction of rows and along the columns of the two-dimensional region for searching for displacements of the television image of the object, the third (56) and fourth (59) buffer random access memory, the imager (57) of the signals of the static and dynamic reference television images of the object, the imager (60) of signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for displacements of the television image of the object, the second node (58) for scaling and shifting the signals of the television image, the switch (62) of the data the minimum values of the signals of the measure of dissimilarity of television images, the approximator (63) of the sequence of the minimum values of the signals of the measure of dissimilarity of television images of a fourth degree polynomial, the node (64) for determining the coordinates of the television image of the object in the analysis window by the minimum position of the approximating polynomial of the fourth degree of the data of the sequence of the minimum values of measure signals dissimilarities of television images, the node (65) determining the coordinates of the television image of the object that in the analysis window on the shift of the boundary of the region of rapid growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window, a node (66) for determining the coordinates of the television image of the object in the analysis window according to the coordinates of the center of the analysis window, the first input of the block (82) for determining the coordinates of the television image object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images is connected to the first inputs of the analyzer (55) update conditions with of the static television reference image of the object and the signal shaper (60) of the measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for the displacements of the television image of the object, the second input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images connected to the first input of the second video switch (53), the third input of the block (82) determining the coordinates of the television from image of the object relative to the center of the current analysis window based on the formation of signals of a measure of dissimilarity of television images is connected to the second input of the second switch (53) of video data and to the first input of a nonlinear high-pass filter (54), the output of which is connected to the third input of the second switch (53) of video data, the fourth input of the block (82) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images connected n with the third input of the signal shaper (60) of the measure of dissimilarity between the signals of the television images in the two-dimensional region of the search for the displacement of the television image of the object and with the second inputs of the shaper (57) of the signals of the static and dynamic reference television images of the object and the analyzer (55) of the conditions for updating the static reference television image object, the fifth input of the block (82) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation the signals of the measure of dissimilarity of television images is connected to the fourth input of the imager (60) of the signals of the measure of dissimilarity between the signals of television images in the two-dimensional region for searching for displacements of the television image of the object, the sixth input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of measure signals dissimilarities of television images is connected to the third input of the shaper (57) of the signals of the static and dynamic reference television images of the object, the seventh input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the fourth input of the shaper (57) of the signals of the static and dynamic reference television images of the object and to the fifth input of the shaper ( 60) signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region of the search for the displacements of the television image of the object, the eighth input d block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the second input of the nonlinear high-pass filter (54), with the fifth input of the analyzer (55) of the update conditions of the static reference television image of the object, the fourth input of the switch (62) data sequence of the minimum values of the signals of the measure of dissimilarity of television images and the seventh input of the shaper (60) signals differences between the signals of the television images in the two-dimensional region for searching for the displacements of the television image of the object, the output of the second switch (53) of the video data is connected to the second input of the shaper (60) of signals for the dissimilarity between the signals of the television images in the two-dimensional region for searching for the displacements of the television image of the object and the input of the third buffer operational a storage device (56), the output of which is connected to the first input of the shaper (57) of the signals of the static and dynamic reference television isionic images of the object, the first output of the analyzer (55) for updating the static reference television image of the object is connected to the fifth input of the shaper (57) of signals of static and dynamic reference television images of the object, the output of which is connected to the input of the second node (58) for scaling and shifting the signals of the television image the output of which is connected to the input of the fourth buffer random access memory (59), the output of which is connected to the sixth input of the signal shaper (60) measures of the dissimilarity between the signals of the television images in the two-dimensional region of searching for the displacements of the television image of the object, the output of which is connected to the input of the node (61) for generating and analyzing the type of sequences of the minimum values of the signals of the measure of the dissimilarity of the television images along the rows and along the columns of the two-dimensional region of the search for the displacements of the television image of the object, the first the output of which is connected to the second output of the block (82) for determining the coordinates of the television image of the object relative to the center of the current o the analysis window based on the generation of signals of the measure of dissimilarity of television images, the third input of the analyzer (55) of the conditions for updating the static reference television image of the object and the input of the switch (62) of the sequence of minimum values of the signals of the measure of dissimilarity of television images, the first output of which is connected to the input of the approximator (63 ) the sequence of minimum values of the signals of the measure of dissimilarity of television images by a polynomial of the fourth degree, the first output of which is connected with the input of the node (64) for determining the coordinates of the television image of the object in the analysis window according to the minimum position of the approximating polynomial of the fourth degree of the data sequence of the minimum values of the signals of the dissimilarity of television images, the second and third outputs of the switch (62) of the data of the sequence of the minimum values of the signals of the dissimilarity of the television images the inputs of the node (65) for determining the coordinates of the television image of the object in the analysis window by the offset of the boundary of the region of rapid growth the signal values of the measure of dissimilarity of the television images relative to the center of the analysis window and the node (66) for determining the coordinates of the television image of the object in the analysis window according to the coordinates of the center of the analysis window, the first outputs of the node (64) for determining the coordinates of the television image of the object in the analysis window for the minimum approximating polynomial fourth the degree of data sequence of the minimum values of the signals of the measure of dissimilarity of television images, node (65) determining the coordinates of the television image of the object in the analysis window by shifting the boundary of the region of rapid growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window and the node (66) for determining the coordinates of the television image of the object in the analysis window by the coordinates of the center of the analysis window are connected to the first output of the block for determining the coordinates of the television image object relative to the center of the current analysis window based on the formation of signals of a measure of dissimilarity of television images and with the fourth input of the analyzer (55) update conditions of the static reference television image of the object, the third output of the block (82) of determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of television images is connected to the second outputs of the node (56) for the formation and analysis of the type of sequences of the minimum values of signals of the measure of dissimilarity of television images along the rows and along the columns of the two-dimensional search region for the displacements of the television image of the object, the analyzer (55 ) conditions for updating a static reference television image of an object, an approximator (63) of a sequence of minimum values of signals of a measure of dissimilarity of television images of a fourth degree polynomial, a node (64) of determining coordinates of a television image of an object in the analysis window by the minimum position of an approximating polynomial of the fourth degree of data of a sequence of minimum values of measure signals dissimilarities of television images, node (65) determining the coordinates of a television image object in the analysis window from the displacement of the boundary of rapid growth dissimilarity measure signal values television image relative to the window center and analysis unit (66) determining the coordinates of the television picture of the object in the analysis window center coordinates of the analysis window.

 The invention is illustrated by drawings, where on figa, 1b - the formation of the standard, the first acting L + 1 frames, figv - distribution of the brightness of the image elements of the standard formed on the first frames, and fig.1g is a typical view of the difference correlation function And? Ν , μ), calculated in accordance with expression (1), leading to a breakdown in tracking, Fig. 1e, 1e is the distribution of the brightness of elements in the image of the reference and the current frame, leading to a breakdown in tracking, Fig. 2 is the location of the windows of the PSO and the background frame in the frame, FIG. 3 is an illustration of the brightness distribution along the image line of a scene with a light object on a dark background or with a dark object on a light background, FIG. 4 - illustration of the brightness distribution along the image line of the scene in which the erroneous classification occurs; FIG. 5 - illustration of the brightness distribution along the image line of the scene when the object enters behind an opaque contrasting background object; FIG. 6 - typical view of the impulse response hx [i] of the field drive view of the video camera of the monitoring system, Fig. 7 - temporal reference of the received signals of controlled movement dx and dy of the axis of the field of view of the video camera of the monitoring system to the timing diagram of receiving and processing fields of The images, 8 - block diagram of a first embodiment of the apparatus determine the coordinates of objects 9 - a block diagram of the block 4, FIG. 10 is a block diagram of block 10, FIG. 11 is a block diagram of block 11, FIG. 12 is a block diagram of shaper 32, FIG. 13 is a block diagram of former 36, FIG. 14 is a block diagram of a second embodiment of an object coordinate determining device, FIG. 15 - structural diagram of block 6, FIG. 16 is a diagram of a switch 13 and 84, FIG. 17 is a diagram of a switch 31 of block 10 and block 81, FIG. 18 is a diagram of a switch 53 of block 11 and block 82, FIG. 19 is a flowchart of the inventive device for determining the coordinates of objects according to the first option, Fig. 20 is a flowchart of the operation of the inventive device according to the second embodiment.

 The inventive method of processing signals to determine the coordinates of objects observed in a sequence of television images provides for the possibility of solving the problem in conditions of both the absence and the presence of sensors of uncontrolled movement and the roll of the field of view of the video camera of the monitoring system, as well as the presence, but insufficient accuracy of these sensors .

 When applying the method in the absence or presence, but insufficient accuracy of the sensors of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system, the inventive method is as follows.

 Before the reception of signals of the current n-th field of the television image, where n = 3, 4, 5, ..., determine the controlled movement dx [n] and dy [n] of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals (n -1) and (n-2) -th fields of the television image horizontally and vertically, respectively, due to the action on the camcorder of the control signals moving its field of view.

где X_УПР[i] , Y_УПР[i] - сигналы управления перемещением оси поля зрения видеокамеры системы наблюдения по горизонтали и вертикали, сформированные в результате обработки i-го поля телевизионного изображения,
hx[i] - импульсная характеристика горизонтального привода поля зрения видеокамеры системы наблюдения, представляющая собой отклик приращения угловых координат оси поля зрения видеокамеры системы наблюдения за время между приемом i-го и (i-1)-гo полей телевизионного изображения на воздействие сигнала управления по горизонтали, постоянного на интервале получения первого поля телевизионного изображения и равного нулю в другие моменты времени,
hy[i] - импульсная характеристика вертикального привода поля зрения видеокамеры системы наблюдения, представляющая собой отклик приращения угловых координат оси поля зрения видеокамеры системы наблюдения за время между приемом i-го и (i-1)-гo полей телевизионного изображения на воздействие сигнала управления по вертикали, постоянного на интервале получения первого поля телевизионного изображения и равного нулю в другие моменты времени,
К - длина импульсной характеристики, представляющая собой число полей телевизионного изображения, по истечении которых модуль импульсной характеристики не превышает заданного уровня.The controlled movement dx [n] and dy [n] of the axis of the field of view of the video camera of the surveillance system is determined, for example, by compiling the convolution of the control signals X _SPS [i], Y _SPS [i] by moving the field of view of the camera of the surveillance system with impulse characteristics hx [i] and hy [i] her drives

where X _UPR [i], Y _UPR [i] - control signals for the movement of the axis of the field of view of the video camera of the surveillance system horizontally and vertically, formed as a result of processing the i-th field of the television image,
hx [i] is the impulse response of the horizontal drive of the field of view of the video camera of the monitoring system, which is the response of the increment of the angular coordinates of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the ith and (i-1) -th fields of the television image on the influence of the control signal horizontal constant in the interval of obtaining the first field of the television image and equal to zero at other points in time,
hy [i] is the impulse response of the vertical drive of the field of view of the video camera of the monitoring system, which is the response of the increment of the angular coordinates of the axis of the field of view of the video camera of the monitoring system during the time between reception of the ith and (i-1) -th fields of the television image on the influence of the control signal vertical constant in the interval of obtaining the first field of the television image and equal to zero at other points in time,
K is the length of the impulse response, which is the number of fields of the television image, after which the impulse response module does not exceed a predetermined level.

The purpose of determining the controlled movement of the axis of the field of view of the video camera of the surveillance system is to take it into account when analyzing changes in the images of objects observed in the moving field of view of the video camera. For small angular displacements of the axis of the field of view of the video camera of the surveillance system horizontally and vertically during the change of frames (fields), the change in the television image of the observed object is reduced to its shift by some values dx [n] and dy [n], depending on previously issued control actions on field-of-view drives of a video camera of a surveillance system. Since control actions are generated by processing discretely changing with the frequency of changing fields of television images, the displacements dx [n], dy [n] depend on n - the number of the current field of the television image. In many fields of application of television surveillance systems, the field of view drives of a video camera of a surveillance system with respect to control signals can be considered linear, that is, those for which the connection of output dx [n], dy [n] and input X _control [i], Y _control [ i] signals are defined by convolution type expressions (39).

The control signals of the drives X _UPR [i], Y _UPR [i] here have the meaning of a set angular velocity of the axis of the field of view of the video camera of the surveillance system, and the impulse responses hx [i] and hy [i] are the responses of the increment of the angular position of the axis of the field of view of the video camera monitoring systems on the influence of control signals in the form of a discrete delta function, taken at discrete time points i * T, multiples of the period T of the change of fields of the television image (see Fig.6).

 A typical view of the impulse response hx [i] of the horizontal field of view of the video camera of a surveillance system is shown in FIG. 6.

 The practically impulse characteristics of the field of view actuators of a video camera of a surveillance system can be determined, for example, as follows.

A small test contrast object is placed in the field of view of the video camera, the coordinates of the television image of which are easily determined. A horizontal control signal X _UPR [i] is supplied to the drive input of the field of view of the video camera of the surveillance system in the form of a discrete delta function, as in FIG. 6a. Within a given number of half-frames, the coordinates of the television image of the test object are determined and stored, changing due to the processing of the control signal by the drive. The increments of the coordinates of the television image of the test object are calculated depending on the number of the field of the television image, counted from the moment the test signal is sent. Measurements are performed several times, the results are averaged. After normalizing the indicated increments of the coordinates of the television image of the test object in amplitude, the desired impulse response of the horizontal drives hx [i] is obtained.

 Similarly, the impulse response of hy [i] drives is obtained vertically.

 The timing of the received signals of controlled movement dx [n] of the axis of the field of view of the video camera of the horizontal monitoring system to the timing diagram of receiving and processing fields of the television image is illustrated in Fig. 7.

 On figa schematically shows the time intervals of obtaining the fields of the television image and the field blanking pulses. The shading also shows the time intervals for receiving television signals from the analysis window.

In FIG. 7b shows the time intervals of processing television signals analysis window, ending the formation of the control signals _CTRL X driven horizontal field of vision video camera surveillance system.

On figv shows the moments of change of the signal X _CRC [t] relative to the intervals of receipt from the camera fields of the television image.

 In FIG. 7d shows the moments of change of the signal dx [t] relative to the intervals of receipt of the fields of the television image from the camera.

 In FIG. 7 shows that the horizontal component dx [n] of the controlled movement of the axis of the field of view of the video camera of the monitoring system during time T between the reception of the signals of the (n-1) and (n-2) th fields of the television image is calculated during the quenching pulse immediately preceding the receipt of the television signals of the current n-th field and, thus, its readiness for use in processing television signals of the current n-th field is ensured.

 Timing diagrams of the change in the vertical component dy [n] of the controlled movement of the axis of the field of view of the video camera of the surveillance system are similar to those discussed above.

 The speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system is determined from the data of the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals from the previous and current fields of the television image.

 The signals of the current field of the television image and the signals of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system are received and stored. These signals, as well as the signals of the controlled movement of the axis of the field of view of the video camera of the surveillance system are used to generate the signals of the predicted coordinates of the television images 2N landmarks in the current field of the television image.

 To do this, select from the signals of the previous field of the television image the signals of the television images 2N landmarks, and N = 3, 4, 5, ...

 Landmarks are selected, for example, as follows.

In the field of the television image (in interlaced scanning - in half frame) of the N _p x M _p format, the valid search area for terrain landmarks with smaller dimensions (N _p -2dx [n]) x (M _p -dy [n]), taking into account the movement dx, is determined [n] and dy [n] fields of view of the video camera of the horizontal and vertical surveillance system per half frame.

где iok и pok - координаты центра окна с номером nok, nok=l,..., NOP,
(2N₀+1), (2M₀+1) - размеры окон.The search area for landmarks is divided into NOP = 100 - 150 adjacent to each other windows in which the value of the SXY criterion is determined:

where iok and pok are the coordinates of the center of the window with the number nok, nok = l, ..., NOP,
(2N ₀ +1), (2M ₀ +1) - window sizes.

 The values of the SXY criterion (nok) are compared with the threshold and those that exceed the threshold are ranked in descending order, after which they form a list of NK window numbers - candidate candidates with the highest values of the SXY (nok) criterion. N pairs of landmarks are selected from the list of NK candidates. Landmarks included in the pair are selected by the criterion of the maximum distance between the landmarks both horizontally and vertically and when these distances exceed the threshold value. The coordinates iok and pok of the center of the corresponding window are taken as the coordinates of the landmarks.

 From the signals of the television images 2N of the landmarks of the current field of the television image and the signals of the television images of the corresponding 2N landmarks in the previous field of the television image, signals of the measure of dissimilarity of the images of 2N landmarks are generated and they are used to determine the bias of the images 2N of landmarks between the reception of signals from the current fields of the television images.

 Using the offset signal of the television image 2N landmarks for each of the N pairs of landmarks determine the parameters of the shift and rotation of the signals of the current field of the television image during the time between receiving signals of the current fields of television images, for example, as follows.

где X_k(n), Y_k(n) и X_k+1(n), Y_k+1(n) - координаты k-гo и (k+1)-гo ориентиров в текущем n-м полукадре,
Δφ - изменение угла крена поля зрения видеокамеры системы наблюдения за время между приемом полей телевизионного изображения,
ΔХ и ΔY - перемещение оси поля зрения видеокамеры системы наблюдения по горизонтали и вертикали за время между приемом полей телевизионного изображения.For each pair of landmarks (with numbers k and k + 1), they first form a system of 4 linear equations
a ₁₁ t1 + a ₁₂ t2 + a ₁₃ t3 + a ₁₄ t4 = X _k (n),
a ₂₁ t1 + a ₂₂ t2 + a ₂₃ t3 + a ₂₄ t4 = Y _k (n),
а ₃₁ t1 + а ₃₂ t2 + а ₃₃ t3 + а ₃₄ t4 = X _{k + 1} (n), (41)
a ₄₁ t1 + a ₄₂ t2 + a ₄₃ t3 + a ₄₄ t4 = Y _{k + 1} (n),
regarding 4 unknown t1-14:

where X _k (n), Y _k (n) and X _{k + 1} (n), Y _{k + 1} (n) are the coordinates of the k-th and (k + 1) -th landmarks in the current nth half-frame,
Δφ is the change in the angle of heel of the field of view of the video camera of the surveillance system during the time between the reception of the fields of the television image,
ΔX and ΔY - moving the axis of the field of view of the camera of the surveillance system horizontally and vertically during the time between receiving fields of the television image.

In the system of linear equations (40), the coefficients a _{i, j} are equal to:
a ₁₁ = X _k (n-1) a ₁₂ = Y _k (n-1) a ₁₃ = 1 a ₁₄ = 0,
a ₂₁ = Y _k (n-1) a ₂₂ = -X _k (n-1) a ₂₃ = 0 and ₂₄ = 1,
a ₃₁ = X _{k + 1} (n-1) a ₃₂ = Y _{k + 1} (n-1) a ₃₃ = 1 a ₃₄ = 0, (43)
a ₄₁ = Y _{k + 1} (n-1) a ₄₂ = -X _{k + 1} (n-1) a ₄₃ = 0 a ₄₄ = 1,
where X _k (n-1), Y _k (n-1) and X _{k + 1} (n-1), Y _{k + 1} (n-1) are the coordinates of the k-th and (k + 1) -th landmarks in the previous (n-1) -th half frame.

 The system (40) of four linear equations with four unknowns t1-t4 is solved, for example, by the Gauss elimination method [A. Glovatskaya Methods and algorithms of computational mathematics. - M .: Radio and communications, 1999 .-- 408 p.: Ill., P. 4-10], after which the change Δφ of the angle of heel of the field of view of the video camera of the surveillance system is determined.

а затем, используя полученные значения t1-t4, образуют систему 2-х линейных уравнений относительно 2-х неизвестных ΔХ и ΔY:

Система (44) двух линейных уравнений с двумя неизвестными ΔХ и ΔY также может решаться методом исключения Гаусса [Гловацкая А.П. Методы и алгоритмы вычислительной математики. - М.: Радио и связь, 1999. - 408 с.: ил., с.4-10] . Решением системы линейных уравнений (44) завершают определение параметров ΔX, ΔY перемещения и Δφ поворота поля зрения видеокамеры системы наблюдения.

and then, using the obtained values of t1-t4, form a system of 2 linear equations for 2 unknown ΔX and ΔY:

System (44) of two linear equations with two unknowns ΔX and ΔY can also be solved by the Gauss elimination method [A. Glovatskaya Methods and algorithms of computational mathematics. - M .: Radio and communications, 1999. - 408 p.: Ill., P. 4-10]. By solving the system of linear equations (44), the determination of the ΔX, ΔY displacement parameters and Δφ rotation of the field of view of the video camera of the observation system is completed.

 The choice of the number N of pairs of landmarks is due to the following considerations. To obtain stable estimates of the parameters of the displacement and the roll of the field of view of the video camera of the surveillance system, it is necessary that the areas of the television image selected as landmarks belong to a fixed background of the terrain. With the automatic selection of landmarks, cases of their erroneous selection are possible, when either foreign moving objects can be selected as the landmarks, or when the carrier of the observation system moves, moving television images of the near background of the terrain. Moreover, estimates of the parameters of the displacement and roll of the field of view of the video camera of the monitoring system obtained from the data of pairs containing erroneous landmarks differ sharply from the rest. Elimination of the influence of erroneous data is achieved as follows.

 The resulting N values of each of the shift and rotation parameters are ranked in ascending or descending order. The k largest and smallest values of the ranked parameters are discarded (k = 0, 1, 2, ...), and the remaining values are averaged.

 An increase in N and k makes it possible to increase the resistance to errors in the choice of landmarks, however, it increases the cost of implementing the determination of the parameters of the bias and the roll of the field of view of the video camera of the surveillance system. Practically acceptable results are achieved at N = 5 and k = l or at N = 7 and k = 2.

 By integrating the current change Δφ of the angle of heel of the field of view of the video camera of the surveillance system, the current angle of heel φ of the field of vision of the camera of the surveillance system is found.

Separate the shift parameters of the signals of the current field of the television image relative to the previous field of the television image relative to the previous field of the television image into the dx, dy components of the controlled displacement of the axis of the field of view of the video camera of the surveillance system and the components r _x , r _{y of} uncontrolled movement axis of the field of view of the video camera of the surveillance system.

 In those conditions of the application of the method where there is no possibility of receiving signals of uncontrolled movement and the roll of the field of view of the video camera of the monitoring system, the sequence of actions differs from the above described only in that the prediction of the coordinates of the television images 2N landmarks in the current field of the television image is not performed. This entails the need to increase the size of the two-dimensional search area of each of the landmarks, which in turn leads to an increase in either the processing time when searching for the position of landmarks in the current field of the television image, or the hardware costs of implementing procedures for finding the position of landmarks.

 The television image signals of the current frame are generated from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the signals of the previous and current fields of the television image and the signals of the current field of the television image with the element-wise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the observation system Nia.

jy - номер строки в кадре, jy=1,...,Мк,
Nк - число элементов телевизионного изображения в строке,
Мк - число строк в кадре телевизионного изображения,
npk - индекс текущего полукадра,
npk = 1 - в нечетных полукадрах,
npk = 0 - в четных полукадрах,

Как видно из изложенного, в рассматриваемой ситуации (отсутствия или недостаточной точности датчиков неуправляемого перемещения и крена поля зрения видеокамеры системы наблюдения) параметры стабилизации телевизионного изображения r_х, r_у и φ определяются и используются один раз за полукадр (являются постоянными в пределах текущего полукадра) и поэтому в этих условиях способ целесообразно применять при наличии медленных дестабилизирующих поле зрения видеокамеры факторов.Moreover, generation of signals L _FRAME (ix, jy) television picture of the current frame with signals from interlaced _P-BLOCK L (i, p, npk) half-frames of the current television image and the television image L _-1KADR (ix, jy) of the previous frame, is carried out e.g. , prediction signal L _fRAME (ix, jy) of the current television picture frame by displacement television image L _1-up (ix, jy) of the previous frame by the amount of managed displacement dx, dy axis of video camera field of view of observation system horizontally and TOWER and accordingly, a time between the receipt of the television image half-frames
L _FRAME (ix, jy) = L _{-1 FRAME} (ix + dx, jy + dy),
and replacing the points of the predicted television image of the current frame with the points of the television image of the current half frame with compensation for uncontrolled displacements r _x , r _y and roll φ of the field of view of the video camera of the surveillance system
L _FRAME [iх (i, p, npk), jy (i, p, npk)] = L _P-FRAME (i, p, npk),
where i is the number of the element in the line of the television image of the current half frame, i = 1, ..., Nk,
ix is the number of the element in the line of the television image of the current frame, ix = 1, ..., Nк,
p is the line number in the half frame,

jy - line number in the frame, jy = 1, ..., Mk,
Nк - the number of elements of the television image in a row,
Mk - the number of lines in the frame of the television image,
npk - index of the current half frame,
npk = 1 - in odd half frames,
npk = 0 - in even half frames,

As can be seen from the above, in the situation under consideration (lack or insufficient accuracy of the sensors of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system), the stabilization parameters of the television image r _x , r _y and φ are determined and used once per half frame (they are constant within the current half frame) and therefore, under these conditions, the method is advisable to apply in the presence of slow factors destabilizing the field of view of the video camera.

 According to the readiness of the signals of the television image in the frame format, the signals of the television image are generated and scaled in the current analysis window using the position and size signals of the current analysis window of the television image and the television image signals of the current frame.

 The signals of the scaled television image are stored in the current analysis window.

 The signals of the differential television image of the moving object selector are generated by subtracting from the signals of the scaled television image in the current analysis window of the signals of the scaled television image previously stored and reduced to the current scale of the television image in the analysis window and shifted by the amount of displacement of the field of view of the video camera of the surveillance system.

 In those conditions of the application of the method, when it is necessary to determine the coordinates of fast moving objects, for example, when documenting take-off and landing maneuvers of aircraft at airports, when creating a differential television image of the selector of moving objects, use the difference of the images of the current and previous frames, and when determining the coordinates of slowly moving objects the difference between the current and stored in K frames of the previously images in the analysis window can be used.

где ix, jy - координаты сигналов отфильтрованного телевизионного изображения S_fil_СДО(ix,jy) относительно центра текущего окна анализа;

NX, NY - размеры текущего окна анализа,
NF, MF - параметры апертуры низкочастотного фильтра по горизонтали и вертикали, соответственно,
di, dj - внутренние переменные низкочастотного фильтра по горизонтали и вертикали, соответственно, di ∈[-NF, NF], dj ∈[-MF, MF],
h_НЧ [di, dj] - импульсная характеристика низкочастотного фильтра,
и формируют из них сигналы L_{2 БИН СДО}(ix,jy) вторичного бинарного телевизионного изображения селектора движущихся объектов
L_{2 БИН СДО}(iх,jу)=1,
если S_fil _СДО(ix,jy)>Porog _СДО 1 и L_{1 БИН СДО} (ix,jy)=1
или S_fil _СДО(ix,jy)>Porog _СДО 0 и L_{1 БИН СДО}(ix,jy)=0,
L_{2 БИН СДО}(iх,jу)=0 - иначе,
где Porog_СДО 1, Porog_СДО 0 - значения порогов принятия решения для единичных и нулевых элементов первичного бинарного телевизионного изображения селектора движущихся объектов соответственно.Generate signals L _{1 BIN LMS} (ix, jy) of the primary binary television image selector moving objects from the signals of the difference television image selector moving objects, for example, by comparing the signals in the module of the difference of television image is adaptively formed threshold subjected signals L _{1 BIN LMS} (ix, jy ) low-pass filtering, for example, using two-dimensional convolution

where ix, jy are the coordinates of the filtered television image S_fil _DLS (ix, jy) relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively,
di, dj are the internal variables of the low-pass filter horizontally and vertically, respectively, di ∈ [-NF, NF], dj ∈ [-MF, MF],
h _LF [di, dj] - impulse response of a low-pass filter,
and form from them signals L _{2 BIN SDO} (ix, jy) of the secondary binary television image of the selector of moving objects
L _{2 BIN SDO} (ikh, jy) = 1,
if S_fil _LMS (ix, jy)> Porog _DLS 1 and L _{1 BIN LMS} (ix, jy) = 1
or S_fil _LMS (ix, jy)> Porog _DLS 0 and L _{1 BIN LMS} (ix, jy) = 0,
L _{2 BIN SDO} (ikh, jy) = 0 - otherwise,
where Porog _SDO 1, Porog _SDO 0 - values of decision thresholds for single and zero elements of the primary binary television image of the selector of moving objects, respectively.

 Simultaneously with storing the signals of the scaled television image in the current analysis window, generating signals of the differential television image of the selector of moving objects and generating signals of the primary binary television image of the selector of moving objects from the signals of the scaled television image in the current analysis window, taking into account the signals of the controlled movement of the field of view of the video camera of the surveillance system form the signals of the primary binary television from mapping the detector background signals and changes the binary classifier television image histogram.

NX, NY - размеры текущего окна анализа,
присвоением значений первичному бинарному телевизионному изображению L_{1 БИН ДИФ}(ix,jy) детектора изменений фона в соответствии с правилом:
L_{1 БИН ДИФ}(iх,jу)=1,
если |L_{p ДИФ}(ix,jy)|≥ПОРОГ_ДИФ(ix,jy),
или
L_{1 БИН ДИФ}(iх,jу)=0,
если |L_{p ДИФ}(ix,jy)|<ПОРОГ_ДИФ(ix,jy),
причем сигналы L_{n эф}(ix,jy) телевизионного изображения эталонного фона формируют разделением сигналов масштабированного телевизионного изображения в текущем окне анализа на сигналы телевизионного изображения трех типов:
сигналы телевизионного изображения в окне объекта - ОО_ДИФ,
сигналы телевизионного изображения в окне фона - ОФ_ДИФ,
сигналы телевизионного изображения в окне - "Новый фон" - НФ,
где в качестве сигналов телевизионного изображения в окне объекта определяют сигналы телевизионного изображения в прямоугольнике, находящемся в центре текущего окна анализа и включающем в себя преимущественно элементы телевизионного изображения объекта, в качестве сигналов телевизионного изображения окна "Новый фон" определяют элементы телевизионного изображения на внешних границах текущего окна анализа, на которых за счет движения объекта и перемещения поля зрения видеокамеры системы наблюдения появляются новые элементы телевизионного изображения фона, в качестве сигналов телевизионного изображения окна фона определяют все оставшиеся элементы телевизионного изображения окна анализа, запоминанием в окне "Новый фон" сигналов L_{n OA}(ix,jy) из текущего окна анализа масштабированного телевизионного изображения текущего n-го кадра:
L_{n эф}(ix,jy)=L_{n OA}(ix,jy), ix,jy∈НФ,
где L_{n OA}(ix, jy) - значение сигнала яркости элемента масштабированного телевизионного изображения в окне анализа с координатами ix,jy,
усреднением в окне фона сигналов масштабированного телевизионного изображения L_{n OA}(ix, jy) из текущего окна анализа рекурсивным фильтром первого порядка с постоянной Wоф и с учетом сдвига окна анализа в инерциальной системе координат за последний кадр:
L_{n эф}(ix,jy)=(1-Wоф)*L_{n-1 эф}(iх+Vx,jу+Vу)+Wоф*L_{n OA}(ix,jy) ix,jy∈ОФ_ДИФ,
где Vx, Vy - перемещение центра окна анализа за последний кадр по горизонтали и вертикали, соответственно, в инерциальной системе координат,
перезаписью в окне объекта сигналов телевизионного изображения эталонного фона предыдущего кадра со сдвигом, учитывающим перемещение центра окна анализа за последний кадр:
L_{n эф}(ix,jy)=L_{n-1 эф}(iх+Vх, jy+Vy) ix,jy ∈ ОО_ДИФ.Moreover, the signals L _{1 BIN DIF} (ix, jу) of the primary binary television image of the background change detector are formed, for example, by converting the signals L _{n 1 eff} (ix, jy) of the television image of the reference background obtained in the previous (n-1) frame to the current the scale, by the formation of signals L _{p DIF} (ix, jy) of the differential television image of the background change detector by subtracting the scaled television image from the signals L _{n OA} (ix, jy) in the current signal analysis window L _{n-1 eff} (ix, jy) of the reference television image background of previous frame a shift that takes into account the movement of Vx, Vy center of the analysis window in the inertial frame of the last coordinates:
L _{p DIF} (ix, jу) = L _{n OA} (ix, jy) - L _{n-1 eff} (ix + Vx, jy + Vy),
determination of the binarization threshold _DIF threshold (ix, jy) of the background change detector as a value proportional to the local scattering parameter of the difference television image L _{r DIF} (ix, jy) background change detector in the vicinity of the point with coordinates ix, jy,
where ix, jy are the coordinates of the signals of the scaled television image relative to the center of the current analysis window;

NX, NY - dimensions of the current analysis window,
assignment of values to the primary binary television image L _{1 BIN DIF} (ix, jy) of the background change detector in accordance with the rule:
L _{1 BIN DIF} (ix, jy) = 1,
if | L _{p DIF} (ix, jy) | ≥POROG _DIF (ix, jy),
or
L _{1 BIN DIF} (ix, jy) = 0,
if | L _{p DIF} (ix, jy) | < _DIF Threshold (ix, jy),
moreover, the signals L _{n eff} (ix, jy) of the television image of the reference background are formed by dividing the signals of the scaled television image in the current analysis window into three types of television image signals:
signals of the television image in the object window - OO _DIF ,
signals of the television image in the background window - OF _DIF ,
signals of the television image in the window - "New background" - NF,
where, as the signals of the television image in the window of the object, the signals of the television image are determined in the rectangle located in the center of the current analysis window and which mainly includes the elements of the television image of the object, the signals of the television image of the "New background" window determine the elements of the television image at the external borders of the current analysis windows where new elements of the TV appear due to the movement of the object and the displacement of the field of view of the video camera of the surveillance system onnogo background image, as background image television signals windows define all remaining elements of the television image analysis windows, latching in the "New background" signals L _{n OA} (ix, jy) of the current window of analysis scaled television image of the current n-th frame:
L _{n eff} (ix, jy) = L _{n OA} (ix, jy), ix, jy∈NF,
where L _{n OA} (ix, jy) is the value of the brightness signal of the element of the scaled television image in the analysis window with the coordinates ix, jy,
averaging in the background window of signals of a scaled television image L _{n OA} (ix, jy) from the current analysis window with a first order recursive filter with constant Wof and taking into account the shift of the analysis window in the inertial coordinate system for the last frame:
L _{n eff} (ix, jy) = (1-Woff) * L _{n-1 eff} (ix + Vx, jу + Vу) + Woff * L _{n OA} (ix, jy) ix, jy∈OF _DIF ,
where Vx, Vy - moving the center of the analysis window for the last frame horizontally and vertically, respectively, in an inertial coordinate system,
overwriting in the window of the object of the signals of the television image the reference background of the previous frame with a shift that takes into account the movement of the center of the analysis window for the last frame:
L _{n eff} (ix, jy) = L _{n-1 eff} (ix + Vx, jy + Vy) ix, jy ∈ OO _DIF .

где NF и MF - параметры апертуры низкочастотного фильтра по горизонтали и вертикали,
h_НЧ[di, dj] - импульсная характеристика низкочастотного фильтра.Low-frequency filtering of signals of a primary binary television image L _{i BIN DIF} (ix, iy) of a detector of background changes can be performed, for example, using two-dimensional convolution

where NF and MF are the aperture parameters of the low-pass filter horizontally and vertically,
h _LF [di, dj] - impulse response of a low-pass filter.

From the signals of the low-pass filter S_fil _DIF (ix, jy) form the signals L _{2 BIN DIF} (ix, jy) of the secondary binary television image of the detector changes the background, for example, in accordance with the rule:
L _{2 BIN DIF} (ix, jy) = 1,
if S_fil _DIF (ix, jy)> Porog _DIF 1 and L _{1 BIN DIF} (ix, jy) = 1
or S_fil _DIF (ix, jy)> Porog _DIF 0 and L _{1 BIN DIF} (ix, jy) = 0,
L _{2 BIN DIF} (ix, jy) = 0 - otherwise,
where Rorog _DIF 1, Rorog _DIF 0 are the decision thresholds for the unit and zero elements of the primary binary television image of the background change detector, respectively.

или
L_{БИН ГК}(ix,jy)=0 - в противном случае,
где L_{n AH}(ix,jy) - сигналы масштабированного телевизионного изображения в текущем окне анализа,
W_n-1 ^H _ЦОО[L_i] - нормированная гистограмма распределения яркостей L_i сигналов масштабированного телевизионного изображения в центральном окне объекта - ЦОО, полученная в предыдущем (n-1)-м кадре,
W_n-1 ^H _ОФГК[L_i] - нормированная гистограмма распределения яркостей сигналов телевизионного изображения фона в окне фона гистограммного классификатора -ОФГК, полученная в предыдущем (n-1)-м кадре,
L_i - уровень яркости сигналов телевизионного изображения,
i - номер уровня яркости сигналов телевизионного изображения, i=1,..., N_yp,
N_yp - число уровней яркости сигналов телевизионного изображения в текущем окне анализа,
α_n-1 - параметр, зависящий от числа элементов телевизионного изображения в окне объекта гистограммного классификатора - ООГК, классифицированных как фон, в предыдущем (n-1)-м кадре, α_n-1 имеет смысл оценки априорной вероятности наличия элемента изображения фона в окне объекта и может определяться, например, в соответствии с формулой

где NN_фон(n-1) - число элементов изображения фона, обнаруженных в центральном окне объекта в (n-1)-м кадре,
A(ix, jy) - штрафная функция, зависящая от координат элемента ix,jy телевизионного изображения в окне анализа, в том числе при ix,jy∈ЦОО A(ix,jy)= A₀, A(ix, jy) - имеет смысл стоимости неправильной классификации элементов телевизионного изображения на элементы объекта и фона, значения устанавливают A(ix, jy) минимальными в центральном окне объекта (А₀) и растущими к краям окна анализа, значение А₀ может быть выбрано, например, в диапазоне значений 1,5,...,2,0.The signals L _{BIN GK} (ix, jy) of the binary television image of the histogram classifier are formed, for example, in accordance with the rule:
L _{BIN GK} (ix, jy) = 1 if

or
L _{BIN GK} (ix, jy) = 0 - otherwise,
where L _{n AH} (ix, jy) are the signals of the scaled television image in the current analysis window,
W _n-1 ^H _COO [L _i ] - normalized histogram of the distribution of brightness L _{i of the} signals of the scaled television image in the central window of the object - COO obtained in the previous (n-1) -th frame,
W _n-1 ^H _OFGK [L _i ] - normalized histogram of the distribution of the brightness of the signals of the television background image in the background window of the histogram classifier-OFGK obtained in the previous (n-1) -th frame,
L _i - the brightness level of the signals of the television image,
i is the number of the brightness level of the television image signals, i = 1, ..., N _yp ,
N _yp is the number of brightness levels of the television image signals in the current analysis window,
α _n-1 is a parameter that depends on the number of elements of the television image in the window of the histogram classifier object - OOGKs classified as background in the previous (n-1) th frame, α _n-1 makes sense of estimating the a priori probability of the presence of a background image element in object window and can be determined, for example, in accordance with the formula

where NN _background (n-1) is the number of background image elements detected in the central window of the object in the (n-1) -th frame,
A (ix, jy) is a penalty function that depends on the coordinates of the element ix, jy of the television image in the analysis window, including at ix, jy∈COO A (ix, jy) = A ₀ , A (ix, jy) - has the meaning of the cost of incorrect classification of television image elements into elements of the object and background, the values are set A (ix, jy) to the minimum in the central window of the object (A ₀ ) and growing towards the edges of the analysis window, the value A ₀ can be selected, for example, in the range of values 1 , 5, ..., 2.0.

After determining the coordinates X _O , Y _O and the dimensions Rxo, Ryo of the television image of the object in the current nth frame, the rectangle with the dimensions Rx _ЦО = Rхо, Ру _Цо = Rуо located inside the analysis window, whose center is coincides with the center X _O , Y _{O of the} television image of the object, and which includes mainly the elements of the television image of the object, as the window of the histogram classifier object - OOGK determine the area enclosed between two rectangles with a common center and dimensions
Rx _COO , Ry _COO and Rx _OOGK , Ry _OOGK ,
where Rx _OOGK > Rx _COO and Ry _OOGK > Ry _COO ,
as the background window of the histogram classifier - OFGK, determine the area enclosed between two rectangles with a common center and with dimensions
Rx _OOGK , Ry _OOGK and Rx _OFGK , Ru _OFGK ,
where Rx _OFGC > Rx _OOGK and Rу _OFGK > Ry _OOGK .

где h_сг[j] - импульсная характеристика сглаживающего фильтра,
(2*ns+1) - число точек импульсной характеристики сглаживающего фильтра,
определяют текущий порог П_г усреднения сглаженной гистограммы

распределения яркостей телевизионного изображения объекта в соответствии с выражением

где П_г0 и k_пг - постоянные величины,
1[х] - единичная функция, определяемая условиями:
1[х]=1 при х≥0,
1[х]=0 при х<0,
усредняют текущий порог Пг рекурсивным фильтром первого порядка

где γ_{n пор} - постоянная фильтра усреднения порога Пг, изменяющаяся в зависимости от номера кадра от значения γ_{1 пор} = 1 в первом кадре до стационарного значения γ_пор,
n - номер текущего кадра телевизионного изображения,
ограничивают значения усредненного порога сверху и снизу,
усредняют сглаженную гистограмму

распределения яркостей L_i телевизионного изображения объекта рекурсивным фильтром первого порядка

где γ_{n ЦОО} - постоянная фильтра усреднения гистограммы распределения яркостей телевизионного изображения объекта, изменяющаяся в зависимости от номера кадра от значения γ_{1 ЦОО} = 1 в первом кадре до стационарного значения γ_ЦОО, причем начиная с номера кадра, превосходящего Nкy,
где Nку - номер кадра, с которого усреднение гистограммы

выполняют по условиям, Nку=16,...,128,
усреднение гистограммы

распределения яркостей L_i телевизионного изображения объекта выполняют только для тех уровней яркости L_i, для которых одновременно выполняются два условия:

В заявляемом способе для повышения статистической устойчивости усредняют только гистограмму распределения яркостей объекта, поскольку изображение объекта в последовательности кадров находится в ЦОО постоянно и относительно медленно изменяется. Использование указанных выше условий для усреднения гистограммы

позволяет снизить вероятность попадания элементов телевизионного изображения фона в гистограмму объекта. Сглаженная гистограмма распределения яркостей фона

не усредняется для того, чтобы при движении объекта в направлении участка фона, имеющего такие же яркости, как и объект, гистограмма

успела бы измениться и не происходила ошибочная классификация элементов телевизионного изображения фона как принадлежащих объекту.From the television image signals L _{n AH} (ix, jy) read from the OFGC window in the current nth frame, the histogram W _OFGC [L _i ] of the brightness distribution L _{i of the} television background image is determined from the television image signals selected from the DSC window in the current n-th frame, determine the histogram W _COO [L _i ] of the brightness distribution L _{i of the} television image of the object, smooth out the histograms W _OFGC [L _i ] and W _COO [L _i ]

where h _cg [j] is the impulse response of the smoothing filter,
(2 * ns + 1) is the number of points of the impulse response of the smoothing filter,
determine the current threshold P _g averaging the smoothed histogram

the brightness distribution of the television image of the object in accordance with the expression

where P _g0 and k _pg are constant values,
1 [x] is the unit function determined by the conditions:
1 [x] = 1 for x≥0,
1 [x] = 0 for x <0,
average the current threshold Pg with a first-order recursive filter

where γ _{n pore} is the filter averaging constant of the threshold Pg, which varies depending on the frame number from the value of γ _{1 pore} = 1 in the first frame to the stationary value of γ _pore ,
n is the number of the current frame of the television image,
limit the values of the averaged threshold above and below,
average the smoothed histogram

the brightness distribution L _{i of the} television image of the object with a first order recursive filter

where γ _{n COO} is the filter averaging constant of the histogram of the brightness distribution of the television image of the object, which varies depending on the frame number from the value of γ _{1 COO} = 1 in the first frame to the stationary value of the _COO , starting from the frame number exceeding Nкy,
where Nk is the frame number from which the histogram averaging

fulfill the conditions, Nku = 16, ..., 128,
histogram averaging

the brightness distribution L _{i of the} television image of the object is performed only for those brightness levels L _i for which two conditions are simultaneously satisfied:

In the inventive method for increasing statistical stability, only the histogram of the brightness distribution of the object is averaged, since the image of the object in the frame sequence is constantly located in the COO and changes relatively slowly. Using the above conditions to average the histogram

allows you to reduce the likelihood of elements of the television image of the background in the histogram of the object. Smoothed histogram of background brightness distribution

not averaged so that when the object moves in the direction of a background section having the same brightness as the object, the histogram

would have time to change and there was no erroneous classification of elements of the television image of the background as belonging to the object.

распределения яркостей L_i телевизионного изображения объекта

где N_yp - число уровней яркости телевизионного изображения L_{n AH}(ix,jy).Form a normalized W _n ^H _COO [L _i ] histogram of the distribution of the brightness of the television image of the object from the average histogram

luminance distribution L _{i of the} television image of the object

where N _yp is the number of brightness levels of the television image L _{n AH} (ix, jy).

распределения яркостей L_i телевизионного изображения фона:

После получения бинарных изображений формируют горизонтальные и вертикальные проекции сигналов вторичных бинарных изображений селектора движущихся объектов и детектора изменений фона и сигналов бинарного телевизионного изображения гистограммного классификатора.A normalized Wn ^H _OFGC [L _i ] histogram of the brightness distribution of the television background image from the smoothed histogram is formed

luminance distribution L _i television background image:

After obtaining the binary images, horizontal and vertical projections of the signals of the secondary binary images of the moving object selector and the detector of changes in the background and signals of the binary television image of the histogram classifier are formed.

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где Nwin и Mwin - размеры вторичного бинарного телевизионного изображения селектора движущихся объектов по горизонтали и вертикали соответственно.Moreover, the horizontal and vertical projections G _{pr SDO} (ix), V _{pr SDO} (jy) signals of the secondary binary television image L _{2BIN sdo} (ix, jy) of the selector of moving objects is determined, for example, in accordance with the expressions

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where Nwin and Mwin are the sizes of the secondary binary television image of the selector of moving objects horizontally and vertically, respectively.

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где Nwin и Mwin - размеры вторичного бинарного телевизионного изображения детектора изменений фона по горизонтали и вертикали соответственно.The horizontal and vertical projections G _{pr DIF} (ix), V _{pr DIF} (jу) of the signals of the secondary binary television image L _{2BIN DIF} (ix, jy) of the detector for changing the background can be determined in accordance with the expressions

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where Nwin and Mwin are the sizes of the secondary binary television image of the background and horizontal changes detector, respectively.

для ix=1,...,Nwin,

для jy=1,...,Mwin,
где Nwin и Mwin - размеры бинарного телевизионного изображения гистограммного классификатора по горизонтали и вертикали соответственно.The horizontal and vertical projections G _{pr GK} (ix), V _{pr GK} (jy) of the binary television image signals L _{2BIN GK} (ix, jy) of the histogram classifier is determined, for example, in accordance with

for ix = 1, ..., Nwin,

for jy = 1, ..., Mwin,
where Nwin and Mwin are the dimensions of the binary television image of the histogram classifier horizontally and vertically, respectively.

причем усредненные плотности

бинарных изображений детектора изменений фона, селектора движущихся объектов и гистограммного классификатора получают в результате ограничения минимальных и максимальных значений и последующего усреднения рекурсивными фильтрами первого порядка текущих плотностей V_ДИФ(n), V_СДО(n), V_ГК(n) соответствующих бинарных изображений,
где

S_ОДИФ(n), S_ОСДО(n), S_ОГК(n) - текущие площади бинарных изображений детектора изменений фона, селектора движущихся объектов и гистограммного классификатора, соответственно, внутри границ телевизионного изображения объекта,
S_ОО(n) - текущая площадь области внутри границ телевизионного изображения объекта,
F_{НУ_ДИФ}(n), F_{НУ_СДО}(n), F_{НУ_ГК}(n) - функции ввода начальных условий детектора изменений фона, селектора движущихся объектов и гистограммного классификатора соответственно,
n - номер текущего кадра.The confidence coefficients W _DIF , W _LMS , W _{GK of} signals of secondary binary images of the detector of changes in the background and the selector of moving objects and binary images of the histogram classifier are determined, for example, as a product of the functions of inputting the initial conditions to the normalized average densities of binary images of the detector of background changes, the selector of moving objects and histogram classifier, respectively

moreover, the averaged densities

binary images of the background change detector, moving object selector, and histogram classifier are obtained by limiting the minimum and maximum values and subsequent averaging of the first densities V _DIF (n), V _SDO (n), V _ГК (n) of the corresponding binary images by recursive first-order filters,
Where

S _ODIF (n), S _OSDO (n), S _OGK (n) - the current area of the binary images of the background change detector, the moving object selector and the histogram classifier, respectively, inside the boundaries of the television image of the object,
S _OO (n) is the current area of the region inside the boundaries of the television image of the object,
F _{NU_DIF} (n), F _{NU_SDO} (n), F _{NU_GK} (n) - the input function of the initial conditions of the detector of changes in the background, the selector of moving objects and the histogram classifier, respectively,
n is the number of the current frame.

Initial conditions input function F _{NU_ DIF} (n), F _{NU_SDO} (n), F _{NU_GK} (n) allow to set as the initial value at the start of automatically determining the coordinate of object confidence factors and to ensure a smooth change them to the normalized average density binary image detector changes in the background, the moving object selector, and the histogram classifier by the time they become statistically stable.

The generalized horizontal G _software (ix, n) and vertical V _software (jy, n) of the projection of the signals of the generalized binary television image of the object are formed from the horizontal and vertical projections of the signals of the secondary binary images of the moving object selector and the background change detector, as well as the signals of the binary images of the histogram classifier based on their joint processing, using reliability coefficients of binary image signals of the histogram classifier, signals of secondary binary images for detector changes and background selector of moving objects, for example, a weighted summation of projections of the binary image changes the background detector, a selector moving object classifier and histogram:
G _prO (ix, n) = W _DIF (n) * G _{pr DIF} (ix) + W _SDO (n) * G _{pr DSS} (ix) + W _GK (n) * G _{pr GK} (ih),
V _prO (jy, n) = W _DIF (n) * V _{pr DIF} (jy) + W _SDO (n) * V _{pr SDO} (jy) + W _GK (n) * V _{pr GK} (jy).

 Horizontal and vertical borders are formed and the dimensions of the television image of the object are determined by the cutoff levels of the left and right set percentages of the area of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object.

 The current and average areas of the generalized binary television image of the object located inside the formed boundaries of the television image of the object are determined.

Determine the current position _OBIN X, Y _OBIN generalized binary television image of the object, for example, as a weighted sum _bct coordinates X, Y _bct generalized center of gravity and _OMED coordinates X, Y _OMED median area television generalized binary object image
X _OBIN (n) = W _CT (n) * X _OSTT + W _MED (n) * X _OMED ,
Y _OBIN (n) = W _CT (n) * Y _OCT + W _MED (n) * Y _OMED ,
W _MED (n) = 1-W _CT (n),
where W _MED (n), W _CT (n) are the weights of the coordinates of the coordinates of the median area and the center of gravity of the generalized binary television image of the object, respectively,
n is the number of the current frame,
moreover, the weight coefficient W _CT (n) of estimating the coordinates of the center of gravity of the generalized binary television image of the object is increased with decreasing average deviation of the object coordinates from their predicted values.

The current velocity of the generalized binary television image of the object in the inertial coordinate system is determined. Moreover, the current horizontal VG _{OB_BIN} and vertical VV _{OB_BIN} components of the estimated speed of the generalized binary television image of the object in the inertial coordinate system can be determined in accordance with the expressions
VG _{OB_BIN} = (dx + ΔX _OA + ΔX _{OB_OA_BIN} ) / T,
VV _{OB_BIN} = (dy + ΔY _OA + ΔY _{OB_OA_BIN} ) / T,
where dx, dy is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively,
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively
ΔX _{OB_OA_BIN} , ΔY _{OB_OA_BIN} - change the coordinates of the binary television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively.

F_НФ[L] - функция двухстороннего ограничения,
F_НФ[L]=L_ПОР при L>L_ПОР,
F_НФ[L]=L при - L_ПОР≤L≤L_ПОР,
F_НФ[L]=-L_ПОР при L<-L_ПОР,
NF, MF - параметры апертуры фильтра по горизонтали и вертикали,
К_НФ, К_ФВЧ - постоянные коэффициенты,
L_ПОР - пороговый уровень функции двухстороннего ограничения.Simultaneously with the processing of signals in the channels of the histogram classification, the detector of changes in the background and the selector of moving objects, and determining on their basis the current coordinates and the speed of the generalized binary television image of the object, the coordinates of the television image of the object relative to the center of the current analysis window are determined on the basis of: Signal formation measures of image dissimilarity as a result of nonlinear high-frequency filtering of signals of the specified television image in the current analysis window, provided that the average area of the generalized binary television image of the object is exceeded by a threshold value, the signals obtained after non-linear filtering of the signal are stored, the signals of the static reference television image of the object or the signals of the static and dynamic reference images of the object are brought in, the signals of the static reference television images of an object or signals of static and dynamic reference images of the object to the current scale, the formation and storage of signals of the measure of dissimilarity between the signals of the images of the current analysis window and the signals of the static or static and dynamic reference images of the object in the two-dimensional search region for the displacements of the television image of the object, determining the minimum values of the signals of the measure of dissimilarity of images along the rows and columns of the two-dimensional areas of searching for displacements of images of an object and the formation of sequences of minimum values of measure signals dissimilarities of the images along the rows and columns of the two-dimensional region of the search for displacements of the television image of the object. Moreover, nonlinear high-frequency filtering of signals of a scaled television image in the current analysis window is performed, for example, in accordance with the expression
L _{n HF} (ix, jy) = L _{n AH} (ix, jy) + K _NF * F _NF [L _{n HPF} (ix, jy)],
Where

F _NF [L] - function of two-way restriction,
F _NF [L] = L _POR for L> L _POR ,
F _NF [L] = L when - L _POR ≤L≤L _POR ,
F _NF [L] = - L _POR for L <-L _POR ,
NF, MF - horizontal and vertical filter aperture parameters,
K _NF , K _HPF - constant coefficients,
L _POR - threshold level of the function of two-way restriction.

 Depending on the restrictions on the allowable reaction time when generating control signals for the position of the axis of the field of view of the video camera of the monitoring system or hardware limitations, the method provides the ability to determine the coordinates of the object in the analysis window based on the processing of the signals of the image dissimilarity measure using either static or static and dynamic reference images of the object.

 Processing using only a static reference television image of an object requires less time or hardware, but is less reliable in determining the conditions for changing a static reference than processing using static and dynamic reference images of an object.

In the case of using only a static television reference image of the object, the signals of the static television reference image of the object are generated, for example, by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with center, coordinate X _{C ET} , Y _{C ET} which is determined by the difference
X _{C ET} = X _{ABOUT PZ} -X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view,
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view,
when the conditions for changing the static reference television image of the object, formed on the basis of the analysis of the parameters of the signals of measures of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the reference television image of the object, as well as the path parameters of the television image of the object, obtained on the basis of the analysis of measure signals dissimilarities.

In the case of using static and dynamic reference images of the object, the signals of the dynamic reference television image of the object are generated, for example, by reading in each frame the signals from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET Whose} Y _{ET ET} is determined by the difference
X _{C ET} = X _{AB PZ} -X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view,
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view,
and the signals of the static reference television image of the object are formed, for example, by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, the coordinates X _{C ET} , Y _{C ET of} which are determined by the difference
X _{C ET} = X _{AB PZ} -X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view,
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view,
when the conditions for changing the static reference television image of the object, formed by comparing the parameters of the signals of the measures of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static and dynamic reference images of the object, as well as comparing the parameters of the trajectories of the television image of the object obtained using static signals and dynamic reference images of the object.

 The coordinates of the television image of the object in the analysis window in the channel for evaluating the image dissimilarity measure are determined depending on the type of sequence of the minimum signal values of the image dissimilarity measure.

 Provided that the sequence is classified as a sequence with two boundaries of areas of rapid growth in the values of the signals of the image dissimilarity measure near the position of its minimum, the coordinates of the television image of the object in the analysis window are determined by analytically approximating the sequence of the minimum values of the signals of the image dissimilarity measure by a fourth degree polynomial and determining the minimum position of the approximating polynomial.

Analytical approximation of the sequence of minimum values of signals of a measure of dissimilarity of images is performed, for example, by the least squares method (see, for example, A. Glovatskaya, Methods and algorithms of computational mathematics. - M .: Radio and communications, 1999. - 408 pp., Ill. , pp. 99-109), leading to the solution of a system of 5 linear equations with respect to 5 unknown coefficients b _i (i = 0, ..., 4) of the approximating polynomial F (x) of the form
F (x) = b ₀ + b ₁ x + b ₂ x ² + b ₃ x ³ + b ₄ x ⁴ .

The minimum of the approximating function is found explicitly by solving the cubic equation
b ₁ + b ₂ x + b ₃ x ² + b ₄ x ³ = 0,
for example, by the trigonometric method (see G. Korn and T. Korn. A Handbook of Mathematics. - Trans. from the second American revised edition. Edited by I. G. Aramovich. M., Nauka, 1973, 832 pp. with ill., pp. 43-44).

 If the sequence can be attributed to the type of sequences with a flat neighborhood of the minimum position and the presence of one region of rapid growth of signal values of the image dissimilarity measure, then the coordinates of the television image of the object in the analysis window are determined as a value proportional to the displacement of the border of the region of rapid growth of signal values of the image dissimilarity measure relative to the center of the window analysis. This makes it possible to increase the stability of tracking over the edges of extended objects of uniform brightness.

 Provided that the sequence is assigned to the type of sequences with a flat neighborhood of the position of the minimum signal values of the image dissimilarity measure in the entire search area for the displacements of the television image of the object, the coordinates of the center of the analysis window are taken for the coordinates of the television image of the object in the analysis window. This eliminates the random movement ("wandering") of the axis of the field of view of the video camera of the surveillance system within the size of the object when tracking extended objects of uniform brightness.

 Using the coordinates of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of images, determine the current speed of movement of the television image of the object in an inertial coordinate system and determine the reliability coefficient of this speed.

Х_ОА+ΔХ_{ОБ_ОА НСХ})/Т,
VV_{ОБ_НСХ}=(dy+ΔY_ОА+ΔY_{ОБ_ОА НСХ})/Т,
где dx, dy - перемещение оси поля зрения видеокамеры за время Т между получением текущего и предыдущего полей телевизионного изображения по горизонтали и вертикали соответственно,
ΔХ_ОА, ΔY_ОА - изменение положения окна анализа в текущем кадре относительно предыдущего кадра по горизонтали и вертикали соответственно,
ΔХ_{ОБ_ОА_НСХ}, ΔY_{ОБ_ОА_НСХ} - изменение координат объекта в окне анализа в текущем кадре относительно предыдущего кадра по горизонтали и вертикали, получаемых на основе формирования сигналов меры несходства изображений.Moreover, the current horizontal VG _{OB_NXX} and vertical VV _{OB_NXX} components of the estimated speed of the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of images, is determined, for example, in accordance with the expressions
VG _{OB_NX} = (dx +

X _OA + ΔH _{OB_OA NSH)} / T,
_{OB_NSKH} VV = (dy + ΔY _{OA OB_OA NSH} + ΔY) / T,
where dx, dy is the movement of the axis of the field of view of the video camera during the time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively
ΔX _{OB_OA_NXX} , ΔY _{OB_OA_NXX} - change the coordinates of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, obtained on the basis of the formation of signals of the measure of dissimilarity of images.

ограничения его максимальных и минимальных значений и усреднения рекурсивным фильтром первого порядка, нормирования полученного усредненного коэффициента сходства

где n - номер текущего кадра,
σ_фmin(n) - минимальное значение среднеквадратического значения сигналов телевизионного изображения фона в М окнах анализа фона,
E_НСХmin - минимальное значение сигналов меры несходства изображений текущего окна анализа и статического эталонного телевизионного изображения объекта в двумерной области поиска смещений телевизионного изображения объекта,

усредненная плотность обобщенного бинарного телевизионного изображения объекта, получаемая в результате определения текущей плотности

обобщенного бинарного телевизионного изображения объекта, ограничения минимальных и максимальных значений текущей плотности V_БИН(n) обобщенного бинарного телевизионного изображения объекта и последующего усреднения рекурсивным фильтром первого порядка ограниченной текущей плотности V_БИН(n) обобщенного бинарного телевизионного изображения объекта,
S_OБИН(n) - текущая площадь обобщенного бинарного телевизионного изображения внутри границ обобщенного бинарного телевизионного изображения объекта,
S_OO(n) - текущая площадь области внутри границ обобщенного бинарного телевизионного изображения объекта.The confidence coefficient W _CX (n) of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of images, is obtained, for example, by determining the current coefficient of similarity

limiting its maximum and minimum values and averaging with a first-order recursive filter, normalizing the obtained average similarity coefficient

where n is the number of the current frame,
σ _fmin (n) is the minimum root mean square value of the background television image signals in M background analysis windows,
E _NSXmin - the minimum value of the signals of the measure of dissimilarity of the images of the current analysis window and the static reference television image of the object in the two-dimensional region of the search for the displacements of the television image of the object,

averaged density of a generalized binary television image of an object obtained by determining the current density

generalized binary television image of the object, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image of the object and then averaging, with a first order recursive filter, the limited current density V _BIN (n) of the generalized binary television image of the object,
S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object,
S _OO (n) is the current area of the region inside the boundaries of the generalized binary television image of the object.

обобщенного бинарного телевизионного изображения, ограничения минимальных и максимальных значений текущей плотности V_бин(n) обобщенного бинарного телевизионного изображения, последующего усреднения ограниченной плотности обобщенного бинарного телевизионного изображения рекурсивным фильтром первого порядка, нормирования усредненной плотности

обобщенного бинарного телевизионного изображения объекта

где S_ОБИН(n) - текущая площадь обобщенного бинарного телевизионного изображения внутри границ обобщенного бинарного телевизионного изображения объекта,
S_ОО(n) - текущая площадь области внутри границ обобщенного бинарного телевизионного изображения объекта,
n - номер текущего кадра,

усредненный коэффициент сходства, получаемый в результате определения текущего коэффициента сходства

ограничения его максимальных и минимальных значений и усреднения рекурсивным фильтром первого порядка,
σ_фmin(n) - минимальное значение среднеквадратического значения сигналов телевизионного изображения фона в М окнах анализа фона,
E_НСХmin - минимальное значение сигналов меры несходства изображений текущего окна анализа и статического эталонного телевизионного изображения объекта в двумерной области поиска смещений телевизионного изображения объекта.The reliability coefficient of the current velocity of the generalized binary television image of the object in the inertial coordinate system is determined. Moreover, the confidence coefficient W _BIN (n) of the current velocity of the generalized binary television image of the object can be obtained, for example, by determining the current density

generalized binary television image, limiting the minimum and maximum values of the current density V _bin (n) of the generalized binary television image, then averaging the limited density of the generalized binary television image with a first order recursive filter, normalizing the average density

generalized binary television image of an object

where S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object,
S _OO (n) is the current area of the region inside the boundaries of the generalized binary television image of the object,
n is the number of the current frame,

average similarity coefficient obtained by determining the current similarity coefficient

restrictions on its maximum and minimum values and averaging by a first-order recursive filter,
σ _fmin (n) is the minimum root mean square value of the background television image signals in M background analysis windows,
E _НСХmin - the minimum value of the signals of the measure of dissimilarity of the images of the current analysis window and the static reference television image of the object in the two-dimensional search region for the displacements of the television image of the object.

A comprehensive assessment of the current speed of moving a television image of an object in an inertial coordinate system is formed from data evaluating the current speed of moving a television image of an object, obtained on the basis of the formation of signals of the measure of dissimilarity of images, and from data evaluating the current speed of moving an object from generalized projections of signals from a generalized binary television image of an object with taking into account the reliability coefficients of the speed components of the television image These and a priori limitations of the object’s maneuvering speed, for example, a complex estimate of the horizontal VG _OB (n) and vertical VV _OB (n) components of the current speed of the television image of the object in the inertial coordinate system is determined by limiting the estimates of the speed of the binary television image of the object and the speed of the television images of the object obtained on the basis of the formation of signals of a measure of dissimilarity of images, the minimum and maximum values formed taking into account m previous values of the complex evaluation of the current speed of movement of a television image of the object, and forming a weighted sum of the count rate of movement limited television binary object image and the speed of movement of television images of the object obtained based on the signal generating dissimilarity measure images
VG _ON (n) = W _BIN (n) * VG _{BIN ON} _ (n) + W _SC (n) * VG _{ON HCX} (n),
VV _ON (n) = W _BIN (n) * VV _{BIN ON} _ (n) + W _SC (n) * VV _{ON HCH} (n),
where W _BIN (n), W _CX (n) are the reliability coefficients of the current velocity of the generalized binary television image of the object and the current velocity of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of images, respectively,
VG _OB _ _BIN (n) and VV _OB _ _BIN (n) - limited horizontal and vertical components of the current speed of movement of the generalized binary television image of the object,
VG _{ABOUT NSX} (n) and VV _{ABOUT NSX} (n) - limited horizontal and vertical components of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of images,
n is the number of the current frame.

 The complex estimate of the current speed of the television image of the object is averaged, for example, using a first-order recursive low-pass filter and the values of the averaged speed are stored.

 The coordinates of the television image of the object in the field of view of the camera are determined by integrating the difference of the complex estimate of the current speed of the television image of the object and the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system in an inertial coordinate system, the initial coordinates of the television image of the object in the field of view of the camera and the start signal of tracking surveillance systems. The coordinates of the television image of the object are extrapolated.

 The control signals for the movement of the axis of the field of view of the video camera of the surveillance system are generated using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system obtained as a result of processing the television image in the current analysis window, or using the extrapolated coordinates and the speed of the television image of the object depending on the results of the analysis of the current and the average speed of the television image of the object, the area and coordinates of the boundaries of the binary The image histogram classifier M windows background analysis, the extrapolated velocity of the television image of the object is formed on the basis of analysis of the stored values of the averaged complex estimation of the speed of movement of television images of the object.

 The position and size signals of the television image analysis window are generated for the next frame using the coordinates of the television image of the object in the field of view of the camera and the dimensions of the television image of the object obtained by processing the current frame, and the initial values of the position and size signals of the television image of the object and the start tracking signal receive from a surveillance system.

 Simultaneously with the processing of signals in the channels of the histogram classifier, the detector of changes in the background, the selector of moving objects and the estimation of the image dissimilarity measure, M background analysis windows are formed along the perimeter of the analysis window and projections of the binary television image signals of the histogram classifier in M background analysis windows are determined, which determine the areas and the coordinates of the boundaries of the binary images of the histogram classifier in M background analysis windows.

площади обобщенного бинарного телевизионного изображения объекта, к примеру, путем проверки выполнения условия

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта или условия,

для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры, полученным в результате обработки телевизионного изображения в текущем окне анализа,
где ks1 - постоянный коэффициент, ks1<1,
ks2(n-n_ЭК) - коэффициент, уменьшающийся с ростом номера кадра n, начиная с номера кадра n_ЭК перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам, ks2(n-n_ЭК)≤ks1.Analyze current S _OBIN (n) and averaged

the area of the generalized binary television image of the object, for example, by checking the condition

to switch to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the extrapolated coordinates of the television image of the object or condition,

to go to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the coordinates of the television image of the object in the field of view of the camera obtained by processing the television image in the current analysis window,
where ks1 is a constant coefficient, ks1 <1,
ks2 (nn _EC ) - a coefficient that decreases with increasing frame number n, starting from frame number n _{EC of the} transition to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates, ks2 (nn _EC ) ≤ks1.

- для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам или выполнения условий

- для перехода к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры, полученным в результате обработки телевизионного изображения в текущем окне анализа,
где VG_ТОБ(n), VV_ТОБ(n) - горизонтальная и вертикальная составляющие текущей скорости перемещения телевизионного изображения объекта, получаемые в результате рекурсивной фильтрации первого порядка с постоянной фильтра 0<WV<1 составляющих VG_ОБ(n), VV_ОБ(n) комплексной оценки скорости перемещения телевизионного изображения объекта
VG_TOБ(n)=VG_TOБ(n-1)+WV1*[VG_OБ(n)-VG_TOБ(n-1)],
VV_TOБ(n)=VV_TOБ(n-1)+WV1*[VV_OБ(n)-VV_TOБ(n-1)],

горизонтальная и вертикальная составляющие усредненной скорости перемещения телевизионного изображения объекта, получаемые в результате рекурсивной фильтрации первого порядка с постоянной фильтра 0<WV2<1 составляющих VG_ОБ(n), VV_ОБ(n) комплексной оценки скорости перемещения телевизионного изображения объекта

причем WV1>WV2.Analysis of the current and average speed of the television image of the object is carried out, for example, by checking the fulfillment of conditions

- to switch to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates or to fulfill the conditions

- to proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the coordinates of the television image of the object in the field of view of the camera, obtained as a result of processing the television image in the current analysis window,
where VG _TOB (n), VV _TOB (n) are the horizontal and vertical components of the current speed of the television image of the object, obtained as a result of first-order recursive filtering with filter constant 0 <WV <1 of the components VG _OB (n), VV _OB (n ) a comprehensive assessment of the speed of movement of the television image of the object
VG _TO (n) = VG _TO (n-1) + WV1 * [VG _TO (n) -VG _TO (n-1)],
VV _TO (n) = VV _TO (n-1) + WV1 * [VV _TO (n) -VV _TO (n-1)],

horizontal and vertical components of the average speed of moving a television image of an object, obtained as a result of first-order recursive filtering with a filter constant of 0 <WV2 <1 of components VG _OB (n), VV _OB (n) of a complex estimate of the speed of moving a television image of an object

moreover, WV1> WV2.

где S_i(n) - значения текущей площади бинарных телевизионных изображений гистограммного классификатора внутри i-гo окна анализа фона, i=1,...,М,

значения усредненной площади бинарных телевизионных изображений гистограммного классификатора внутри i-гo окна анализа фона, i=1,...,М, получаемые на выходе рекурсивного фильтра первого порядка с постоянной фильтра 0<WS<1:

Sпорог(n) - порог обнаружения изменения площади бинарных телевизионных изображений гистограммного классификатора внутри М окон анализа фона

где k - постоянный коэффициент,

усредненная площадь обобщенного бинарного телевизионного изображения объекта, получаемая на выходе рекурсивного фильтра первого порядка с постоянной фильтра 0<WS_ОБ<1:

n - номер текущего кадра,

отношение объект/фон,
где

среднее значение сигналов телевизионного изображения в окне объекта,

среднее значение сигналов телевизионного изображения в окне фона,
σ_Ф - среднеквадратическое отклонение сигналов телевизионного изображения в окне фона от

F_ОГР(Q) - функция ограничения минимальных и максимальных значений,
при выполнении указанных выше условий для текущей площади бинарных телевизионных изображений гистограммного классификатора в одном или нескольких окнах анализа фона с номерами ki,
где ki - номера окон анализа фона, в которых обнаружено изменение текущей площади бинарного телевизионного изображения гистограммного классификатора, ki=1,...,кN,
кN - число окон анализа фона, в которых обнаружено изменение текущей площади бинарного телевизионного изображения гистограммного классификатора, кN≤М, в этих окнах анализируют координаты двух взаимно перпендикулярных границ бинарного телевизионного изображения гистограммного классификатора, расположенных со стороны окна объекта, причем в окнах анализа фона, размещенных со стороны вертикальных границ окна объекта, анализируют перемещение вертикальной границы бинарного телевизионного изображения по горизонтали и контролируют принадлежность координат горизонтальной границы вертикальным координатам окна объекта, а в окнах анализа фона, размещенных со стороны горизонтальных границ окна объекта, анализируют перемещение горизонтальной границы бинарного телевизионного изображения по вертикали и контролируют принадлежность координат вертикальной границы горизонтальным координатам окна объекта, при этом анализ перемещения границ бинарных телевизионных изображений в М окнах анализа фона проводят посредством проверки условий нахождения соответствующих границ бинарного телевизионного изображения гистограммного классификатора внутри границ внутренней O_ki1 и внешней O_ki2 областей ki-гo окна анализа фона с формированием признаков П_ki1 и П_ki2 принадлежности границ бинарных телевизионных изображений к внутренней O_ki1 и к внешней O_ki2 областям ki-гo окна анализа фона,
где П_ki1= 1, если граница бинарного телевизионного изображения гистограммного классификатора лежит внутри границ области О_ki1, или П_ki1=0 - в противном случае,
П_ki2=1, если граница бинарного телевизионного изображения гистограммного классификатора лежит внутри границ области O_ki2, или П_ki2=0 - в противном случае,
причем при обнаружении переходов соответствующих границ бинарного телевизионного изображения гистограммного классификатора от внешней области к внутренней области окон анализа фона с номерами ki=kip1,
где kip1 - номера окон анализа фона, в которых обнаружены переходы границ бинарного телевизионного изображения гистограммного классификатора из внешних областей O_kip12 к внутренним областям O_kip11 окон анализа фона, kip1=1,2,...кN,
то есть при последовательном формировании признаков П_kip12=1 при n_kip1= n_kip12, а затем П_kip11=1 при n_kip1=n_kip11, где n_kip11>n_kip12, n_kip11 - номер кадра, в котором в kip1-м окне анализа фона установлено состояние признака П_kip11= 1, n_kip12 - номер кадра, в котором в kip1-м окне анализа фона установлено состояние признака П_kip12=1, и выполнении условия принадлежности координаты второй контролируемой границы бинарного телевизионного изображения гистограммного классификатора в kip1-м окне анализа фона диапазону координат окна объекта, выполняют переход к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта,
присваивают счетчику SP-переходов границ бинарного телевизионного изображения гистограммного классификатора между внутренними и внешними областями окон анализа фона значение SP= 1 при первом обнаружении перехода границы бинарного телевизионного изображения гистограммного классификатора из внешней области O_kip12 к внутренней области O_kip11 или увеличивают счетчик SP на единицу при повторном обнаружении переходов границы бинарного телевизионного изображения гистограммного классификатора из внешних областей O_kip12 к внутренним областям O_kip11 в течение формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта, устанавливают состояния признаков П_kip11=0, П_kip12=0,
устанавливают состояния признаков П_j2=0 в окнах анализа фона с номерами j≠kipl, j=1...кN, и процесс анализа границ бинарного телевизионного изображения гистограммного классификатора начинают сначала,
при обнаружении переходов соответствующих границ бинарного телевизионного изображения гистограммного классификатора из внутренних областей O_kip21 к внешним областям O_kip22 окон анализа фона,
где kip2 - номера окон анализа фона, в которых обнаружен переход границы бинарного телевизионного изображения гистограммного классификатора из внутренних областей O_kip22 к внешних областям О_kip21 окон анализа фона, kip2= 1,2,...,кN,
в течение формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам телевизионного изображения объекта, то есть при последовательном формировании признаков П_kip21=1 при n_kip2=n_kip21, а затем П_kip22=1 при n_kip2=n_kip22 в одном или нескольких окнах анализа фона с номерами kip2, где n_kip22>n_kip21, уменьшают счетчик SP-переходов границ бинарного телевизионного изображения гистограммного классификатора между внутренними и внешними областями окон анализа фона на единицу и, если SP=0, выполняют переход к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа, устанавливают состояния признаков П_kip21=0, П_kip22=0, устанавливают состояния признаков П_j2= 0 в окнах анализа фона с номерами j≠kip2, j= 1...кN, и процесс анализа границ бинарного телевизионного изображения гистограммного классификатора начинают сначала, контролируют время формирования сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по экстраполированным координатам и время нахождения признаков П_i1, П_i2, i=1,...,M, в состоянии П_i1=1, П_i2=1 и при превышении установленных интервалов времени переводят соответствующие признаки в нулевое состояние, переходят к формированию сигналов управления перемещением оси поля зрения видеокамеры системы наблюдения по координатам телевизионного изображения объекта в поле зрения видеокамеры системы наблюдения, полученным в результате обработки телевизионного изображения в текущем окне анализа.The area and coordinates of the boundaries of the binary television images of the histogram classifier are analyzed in M background analysis windows by checking that the conditions for detecting changes in the area of the binary television images of the histogram classifier inside M background analysis windows are met:

where S _i (n) are the values of the current area of binary television images of the histogram classifier inside the i-th window of the background analysis, i = 1, ..., M,

values of the average area of binary television images of the histogram classifier inside the ith window of the background analysis, i = 1, ..., M, obtained at the output of the first order recursive filter with a filter constant of 0 <WS <1:

S threshold (n) - threshold for detecting changes in the area of binary television images of a histogram classifier inside M windows of background analysis

where k is a constant coefficient,

the average area of the generalized binary television image of the object obtained at the output of a first-order recursive filter with a filter constant of 0 <WS _OB <1:

n is the number of the current frame,

object / background ratio,
Where

the average value of the television image signals in the object window,

the average value of the television image signals in the background window,
σ _f - the standard deviation of the signals of the television image in the background window from

F _OGR (Q) - the function of limiting the minimum and maximum values,
when the above conditions are met for the current area of binary television images of the histogram classifier in one or more background analysis windows with numbers ki,
where ki are the numbers of the background analysis windows in which a change in the current area of the binary television image of the histogram classifier was detected, ki = 1, ..., kN,
kN is the number of background analysis windows in which a change in the current area of the binary television image of the histogram classifier was detected, kN≤M, in these windows the coordinates of two mutually perpendicular boundaries of the binary television image of the histogram classifier are located on the side of the object window, and in the background analysis windows, placed from the side of the vertical borders of the window of the object, analyze the horizontal movement of the vertical border of the binary television image and control the the accuracy of the coordinates of the horizontal border to the vertical coordinates of the window of the object, and in the background analysis windows located on the side of the horizontal borders of the window of the object, analyze the vertical horizontal displacement of the binary television image and control the belonging of the coordinates of the vertical border to the horizontal coordinates of the object window, while analyzing the movement of the boundaries of the binary television images in M background analysis windows are carried out by checking the conditions for finding the corresponding bin boundaries of a bar television image of the histogram classifier inside the boundaries of the internal O _ki 1 and external O _ki 2 regions of the ki-th background analysis window with the formation of signs P _ki 1 and P _ki 2 that the boundaries of the binary television images belong to the internal O _ki 1 and to the external O _ki 2 regions ki-go background analysis windows,
where П _ki 1 = 1, if the boundary of the binary television image of the histogram classifier lies inside the boundaries of the region О _ki 1, or П _ki 1 = 0 - otherwise,
П _ki 2 = 1, if the boundary of the binary television image of the histogram classifier lies inside the boundaries of the region O _ki 2, or П _ki 2 = 0 - otherwise,
moreover, when detecting transitions of the corresponding boundaries of the binary television image of the histogram classifier from the outer region to the inner region of the background analysis windows with numbers ki = kip1,
where kip1 are the numbers of the background analysis windows in which transitions of the boundaries of the binary television image of the histogram classifier from the external regions O _kip1 2 to the internal regions O _kip1 1 of the background analysis windows, kip1 = 1,2, ... кN,
that is, when the signs are formed sequentially, П _kip1 2 = 1 for n _kip1 = n _kip1 2, and then П _kip1 1 = 1 for n _kip1 = n _kip1 1, where n _kip1 1> n _kip1 2, n _kip1 1 is the frame number, in which in the kip1st window of the background analysis the status of the sign П П _kip1 1 = 1 is set, n _kip1 2 is the frame number in which in the kip1-th window of the analysis of the background the state of the sign П П _kip1 2 = 1 is set, and the condition of the second controlled the boundaries of the binary television image of the histogram classifier in the kip1st window of the background analysis to the coordinate range of the object window, go to the formation of the control movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates of the television image of the object,
set the counter of SP transitions of the boundaries of the binary television image of the histogram classifier between the internal and external regions of the background analysis windows to SP = 1 when the transition of the boundary of the binary television image of the histogram classifier from the outer region O _kip1 2 to the inner region O _kip1 1 is first _detected or increase the counter SP by unit upon repeated detection of transitions of the boundary of the binary television image of the histogram classifier from the outer regions O _kip1 2 to the inner regions O _{ki p1} 1 during the formation of the control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the extrapolated coordinates of the television image of the object, set the status of the signs П _kip1 1 = 0, П _kip1 2 = 0,
establish the state of signs П _j 2 = 0 in the background analysis windows with numbers j ≠ kipl, j = 1 ... kN, and the process of analyzing the boundaries of the binary television image of the histogram classifier starts again,
upon detecting transitions of the corresponding boundaries of the binary television image of the histogram classifier from the inner regions O _kip2 1 to the outer regions O _kip2 2 of the background analysis windows,
where kip2 are the numbers of the background analysis windows in which the transition of the boundary of the binary television image of the histogram classifier from the inner regions O _kip2 2 to the outer regions O _kip2 1 of the background analysis windows, kip2 = 1,2, ..., kN,
during the formation of the control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the extrapolated coordinates of the television image of the object, that is, with the sequential formation of signs П _kip2 1 = 1 for n _kip2 = n _kip2 1, and then П _kip2 2 = 1 for n _kip2 = n _kip2 2 in one or more background analysis windows with numbers kip2, where n _kip2 2> n _kip2 1, reduce the counter of SP transitions of the boundaries of the binary television image of the histogram classifier between the internal and external regions of the background analysis windows by one and, if SP = 0, doing t the transition to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained as a result of processing the television image in the current analysis window, sets the status of signs П _kip2 1 = 0, П _kip2 2 = 0 , the states of signs П _j 2 = 0 are established in the background analysis windows with numbers j ≠ kip2, j = 1 ... кN, and the process of analyzing the boundaries of the binary television image of the histogram classifier begins again, The time for generating the control signals for the movement of the axis of the field of view of the video camera of the surveillance system according to the extrapolated coordinates and the time for finding the signs P _i 1, P _i 2, i = 1, ..., M, in the state P _i 1 = 1, P _i 2 = 1 and if the specified time intervals are exceeded, the corresponding signs are converted to the zero state, proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system, obtained as a result Follow the processing of the television image in the current analysis window.

 When applying the proposed method in the presence and sufficient accuracy of sensors of uncontrolled movement and roll of the field of view of the video camera of the surveillance system, the composition and sequence of actions of the method differ from the above actions only in the part that precedes the formation and scaling of the television image signals in the analysis window and which consists in the following.

 Before the reception of signals of the current n-th field of the television image, where n = 3, 4, 5, ..., determine the controlled movement dx [n] and dy [n] of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals (n -1) -th and (n-2) -th fields of the television image horizontally and vertically, respectively, due to the action on the video camera of the monitoring system of the signals to control the movement of its field of view.

The controlled movement dx [n] and dy [n] of the axis of the field of view of the video camera of the surveillance system is determined, for example, by compiling the convolution of the control signals X _CRC [i], Y _CRC [i] by moving the field of view of the camera of the surveillance system with impulse characteristics hx [i] and hy [i] of its drives in accordance with expression (39).

 The speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system is determined from the data of the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals from the previous and current fields of the television image.

 The signals of the current field of the television image and the signals of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system are received and stored and used in the formation of the current frame of the television image as follows.

 The television image signals of the current frame are generated from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the signals of the previous and current fields of the television image and the signals of the current field of the television image with the element-wise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the observation system Nia.

jy - номер строки в кадре, jy=1,...,Мк,
Nк - число элементов телевизионного изображения в строке,
Мк - число строк в кадре телевизионного изображения,
npk - индекс текущего полукадра,
npk=1 - в нечетных полукадрах,
npk = 0 - в четных полукадрах,

В рассматриваемой ситуации (достаточной точности датчиков неуправляемого перемещения и крена поля зрения видеокамеры системы наблюдения) параметры стабилизации телевизионного изображения r_х(р), r_у(р) и φ(р), получаемые от датчиков, могут изменяться в пределах текущего полукадра, что выражается в зависимости этих параметров от номера строки р, и поэтому в этих условиях способ может применяться даже при наличии широкополосных дестабилизирующих поле зрения видеокамеры системы наблюдения факторов.Moreover, the formation of signals L _FRAME (ix, jy) of the television image of the current frame during interlacing from signals L _P-FRAME (i, p, npk) of the current half-frame of the television image and television image L _-1CADR (ix, jy) of the previous frame is carried out, for example , predicting the signals L _FRAME (ix, jy) of the television image of the current frame by shifting the television image L _1CADR (ix, jy) of the previous frame by the amount of controlled movement dx, dy of the horizontal and vertical axis of the field of view of the video camera of the surveillance system and, accordingly, during the time between receiving half-frames of a television image
L _FRAME (ix, jy) = L _{-1 FRAME} (ix + dx, jy + dy)
and replacing the points of the predicted television image of the current frame by the points of the television image of the current half frame with compensation for uncontrolled displacements r _x (p), r _y (p) and roll φ (p) of the field of view of the video camera of the surveillance system
L _FRAME [ix (i, p, npk), jy (i, p, npk)] = L _P-FRAME (i, p, npk),
where i is the number of the element in the line of the television image of the current half frame, i = 1, ..., Nk,
ix is the number of the element in the line of the television image of the current frame, ix = 1, ..., Nк,
p is the line number in the half frame,

jy - line number in the frame, jy = 1, ..., Mk,
Nк - the number of elements of the television image in a row,
Mk - the number of lines in the frame of the television image,
npk - index of the current half frame,
npk = 1 - in odd half frames,
npk = 0 - in even half frames,

In the situation under consideration (sufficient accuracy of the sensors of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system), the stabilization parameters of the television image r _x (p), r _y (p) and φ (p) received from the sensors can vary within the current half frame, which is expressed in the dependence of these parameters on the line number p, and therefore, under these conditions, the method can be applied even in the presence of broadband factors destabilizing the field of view of the video camera of the surveillance system.

 The method has a certain versatility with respect to the composition of information about destabilizing factors and provides adaptive settings based on the analysis of available data on these factors, while maintaining operability in the absence of them.

 For example, in the absence of information from any of the signal sensors, on the basis of which the parameters for stabilizing the television image are formed, the method retains the ability to determine the coordinates and hold objects in the center of the field of view of the video camera of the surveillance system, but with corresponding degradation of parameters. So, in the absence of information from the roll sensor, the television image is stabilized in terms of compensation for uncontrolled movements of the field of view of the video camera of the surveillance system horizontally and vertically, and the value of the roll angle of the field of view of the video camera of the surveillance system is taken equal to φ (p) = 0, which in most cases turns out to be sufficient to automatically determine the coordinates and hold the television image of the object in the center of the field of view of the video camera of the surveillance system with the dimensions of the television image eniya object that do not exceed 1/4 of the width of the field of view of systems surveillance cameras.

 The construction of a set of actions of the method in the form of 4 groups of parallel actions, the results of which are shared and mutually complementary, allows for continuous automatic determination of the coordinates and retention of the television image of the object in the center of the field of view of the video camera of the surveillance system even with a sharp change in operating conditions.

 For example, if the sensors of uncontrolled displacements and (or) the roll of the field of view of the video camera of the surveillance system fail, the data of the functional channels (moving object selector and background change detector) that require stabilization of the television image become almost unreliable if the indicated field perturbations of the video camera of the surveillance system in contrast background terrain. According to the method, the parameters of these functional channels are adapted in such a way that the density of the binary images formed by them under the indicated conditions is significantly reduced, resulting in a decrease in their confidence factors, which in turn is equivalent to "self-switching off" of invalid channels. The remaining functional channels (histogram classifier channel and image dissimilarity analysis channel) under many observation conditions are able to continuously determine the coordinates and hold the television image of the object in the center of the field of view of the video camera of the surveillance system.

 In this way, the continuous functioning of the system as a whole and under other adverse changes in the observation conditions is ensured.

 To use the invention in industry, the device according to the first embodiment can be performed, for example, as follows.

 The first input 1.1 of block 1 of the signal processing device for determining the coordinates of objects observed in a sequence of television images (see Fig. 8) is connected to the first input of the device, and the second input 1.2 to the second input of the device. The input of block 2 is connected to the third input of the device. The second output 1-2 of block 1 is connected to the third input 3.3 of block 3, the first output 3-1 of which is connected to the second input 4.2 of block 4, the first output 4-1 of which is connected to the first input 5.1 of block 5, the first 5-1 and second 5 -2 outputs of which are connected to the first 6.1 and second 6.2 inputs of block 6. The second input 5.2 of block 5 is connected to the second output of 1-2 blocks 1. The first output 1-1 of block 1 is connected to the first input 4.1 of block 4 and the first input 7.1 of driver 7 , the second 7.2 and third 7.3 inputs of which are connected to the first 6-1 and second 6-2 outputs of block 6, respectively. The output of block 2 is connected to the third inputs 4.3 and 6.3 of block 4 and block 6, respectively, the fourth input of which 6.4 is connected to the first output 3-1 of block 3. The fourth input 4.4 of block 4 is connected to the second output 1-2 of block 1. The output of the shaper 7 is connected with the first input 8.1 of block 8, and the fourth input 7.4 of driver 7 is connected to the second output 1-2 of block 1, the fifth input 8.5 of block 8, the second input 9.2 of block 9, the third input 10.3 of block 10, the eighth input 11.8 of block 11, the fourth input of 12.4 block 12, the fifth input 13.5 of the switch 13, the fourth input 14.4 of the block 14, the fourth input of the block 15.4 15, the second input 16.2 of block 16, the fourth input 17.4 of block 17, the fifth input 18.5 of block 18, and the third input of block 19.3. 19, sixth input 20.6 of analyzer 20, third input 21.3 of block 21, third input 22.3 of block 22 and sixth input 23.6 of driver 23. The output of block 8 is connected to the seventh input 10.7 of block 10 and with the third input 11.3 of block 11. First output 11-1 of block 11 is connected to the second input 12.2 of block 12. The fourth and fifth inputs of the device are connected to the first 13.1 and second 13.2 inputs of the switch 13, respectively, the first 13-1 and second 13-2 of which outputs are connected to the third 8.3 and second 8.2 inputs of block 8, respectively, and also with the first 10.1 and second 10.2 and the seventh 11.7 and sixth 11.6 inputs of block 10 and bl ka 11, respectively. The sixth input of the device is connected to the second input 3.2 of block 3, and the seventh input is connected to the fifth input 6.5 of block 6. The first input 9.1 of block 9 is connected to the first output 10-1 of block 10, the fifth input of 10.5 of which is connected to the second output 6-2 of block 6 , the first input 11.1 of block 11 and the first input 12.1 of block 12. The second output 10-2 of block 10 is connected to the third input 13.3 of switch 13 and the first input 14.1 of block 14, the second input 14.2 of which is connected to the fourth output 10-4 of block 10, and the third input 14.3 - with the second output 15-2 of block 15, the first input of which 15.1 is connected to the second output 11-2 of block 11, and the second stroke 15.2 - with the second output 14-2 of block 14. The output of block 9 is connected to the first input 16.1 of block 16. The third output 10-3 of block 10 is connected to the first input 17.1 of block 17, the first output 17-1 of which is connected to the fourth input of block 18.4 18, the first 18.1 and second 18.2 inputs of which are connected to the first outputs 14-1 and 15-1 of blocks 14 and 15, respectively, and the third input 18.3 is connected to the first output 12-1 of block 12. The second input 17.2 of block 17 and the first input 19.1 of the block 19 are combined and connected to the second output 6-2 of the unit 6. The first output 16-1 of the unit 16 is connected to the third input 20.3 of the analyzer 20, the second input 20.2 to which is connected to the fourth output 10-4 of block 10, and the first input 20.1 to the fifth output 10-5 of block 10, the sixth output of 10-6 of which is connected to the third input 15.3 of block 15. The second output 20-2 of the analyzer 20 is connected to the second input 11.2 of block 11. The first output 18-1 of block 18 is connected to the fifth input 11.5 of block 11, the second input 19.2 of block 19, the fourth input 20.4 of analyzer 20 and the first input 21.1 of block 21, the first output 21-1 of which is connected to the second input 22.2 of block 22 and the fifth input 20.5 of the analyzer 20, the first output 20-1 of which is connected to the fourth input 10.4 of block 10, the second input of 21.2 block and 21 and the second input 23.2 of the shaper 23, the first input 23.1 of which is connected to the second output 10-2 of the block 10. The first output 19-1 of the block 19 is connected to the fourth input 13.4 of the switch 13, the first input 22.1 of the block 22 and the third input 23.3 of the shaper 23, the fourth input 23.4 of which is connected to the first output 22-1 of block 22, and the fifth input 23.5 is connected to the second output 21.2 of block 21. The first output 23-1 of driver 23 is connected to the output of the device and the first input 3.1 of block 3, and the second output 23-2 - with the fourth input 8.4 of block 8, the sixth input 10.6 of block 10, the fourth input 11.4 of block 11 and a third E inputs 12.3 and 17.3 of the blocks 12 and 17, respectively. The third input 1.3 of block 1 is connected to the fourth output 20-4 of the analyzer 20, and the fourth input 1.4 of block 1 is connected to the seventh output 10-7 of block 10. The output of processor 24 is connected to the second outputs 3-2 of block 3, 4-2 of block 4, 12-2 and 17-2 of blocks 12 and 17, 16-2 of blocks 16, 18-2, 19-2 and 22-2 of blocks 18, 19 and 22, by the third outputs 14-3 and 15-3 of blocks 14 and 15, 20-3 - analyzer 20, 21-3 - block 21 and 23-3 of the shaper 23 and the eighth output 10-8 of block 10 through a bi-directional bus.

 The first input 4.1 of block 4 is connected to the first input 25.1 of the selector 25 (see Fig. 9), the first 25-1, second 25-2 and third 25-3 outputs of which are connected to the first inputs 26.1 of the first, 27.1 of the second and 28.1 of the third calculators 26 , 27 and 28, respectively, the second 26.2, 27.2 and 28.2, the third 26.3, 27.3 and 28.3 and the fourth 26.4, 27.4 and 28.4 whose inputs are connected to the first 4.1, second 4.2 and third 4.3 inputs of block 4, respectively. The first outputs 26-1 of the first 26, 27-2 of the second 27 and 28-2 of the third 28 calculators are connected to the first 29.1, second 29.2 and third 29.3 inputs of the calculator 29, respectively, the first output 29-1 of which is the first 4-1 output of block 4. The fourth input 4.4 of block 4 is connected to the second input 25.2 of the selector 25. The fourth output 25-4 of the selector 25 and the second outputs 26-2, 27-2, 28-2 and 29-2 of the calculators 26, 27, 28 and 29 are combined and connected to the second output 4-2 of block 4.

 The output of the first buffer random access memory (BOS) 30 of block 10 is connected to the second input 31.2 of the switch 31 (see Fig. 10), the first 31-1 and second 31-2 of which are connected to the third 32.3 and sixth 32.6 inputs of the first driver 32, the fourth input 32.4 of which is connected to the fourth input 10.4 of block 10. The first output 32-1 of the driver 32 is connected to the first input 33.1 of the driver 33, the output of which is connected to the first input 34.1 of the block 34. The fourth input is 35.4 of the block. 35 is connected to the second input 31.2 of the switch 31, and the second input 35.2 to the third 36.3 and 32.2 second inputs of the drivers 36 and 32, respectively, and to the second input 10.2 of the block 10. The second input 36.2 of the driver 36 is connected to the third input 35.3 of the block 35, the fifth input 32.5 shaper 32 and the sixth input 10.6 of block 10, and the output with the input of the filter 37, the output of which is connected to the input of the shaper 38, the output of which is connected to the input of the shaper 39, the output of which is connected to the first input 40.1 of the block 40. The output of the BOZU 41 is connected to the third input 42.3 node 42, the second input of which 42.2 with is single with the first input 36.1 of the shaper 36 and with the fifth input 10.5 of the block 10, the first input 42.1 is connected to the first input 35.1 of the block 35, the fourth 36.4 and the first 32.1 inputs of the shapers 36 and 32, respectively, and with the first input 10.1 of the block 10, and the output of the node 42 is connected with the input of the BOSE 43, the output of which is connected to the second input 44.2 node. 44, the first input 44.1 of which is connected to the seventh input 10.7 of block 10, with the inputs of the BOZU 41 and 30, respectively, the fifth input 36.5 of the shaper 36 and the first input 31.1 of the switch 31, and the output is connected to the input of the shaper 45, the output of which is connected to the input of the filter 46, the output of which is connected to the input of the shaper 47, the output of which is connected to the input of the shaper 48, the output of which is connected to the first input 49.1 of block 49, the first output 49-1 of which is connected to the fifth input 50.5 of the analyzer 50, the fourth input 50.4 of which is connected to the first output 40- 1 block 40, third input 50.3 - with the output of the second driver 39, the second input 50.2 - with the first output 34-1 of block 34, and the first 50.1 - with its first input 34.1. The sixth input.50.6 of the shaper 50 is connected to the output of the shaper 48, and the first output 50-1 is connected to the first input 51.1 of the block 51 and to the fourth output 10-4 of the processing unit 10, the fifth output 10-5 of which is connected to the first output 35-1 of the block 35, the first output 10-1 - with the first input 33.1 of the shaper 33, the second output 10-2 - with the first output 51-1 of the block 51, the third output 10-3 - with the second output 51-2 of the block 51, the sixth output 10-6 - with the third output 35-3 of the block 35, and the seventh 10-7 - with the first output 52-1 of the block 52, the input of which is connected to the third output 51-3 of the block 51. The third input 10.3 of the block 10 is connected to the third the input 31.3 of the switch 31, the seventh input 32.7 of the shaper 32, the second input 33.2 of the shaper 33, the second input 34.2 of the block 34, the fifth input 35.5 of the block 35, the sixth input 36.6 of the shaper 36, the second input 40.2 of the block 40, the fourth input 42.4 of the node 42, the second input 49.2 block 49, the seventh input 50.7 of the shaper 50 and the second input 51.2 of the block 51. The eighth output 10-8 of the block 10 is connected to the second outputs 32-2 of the shaper 32, 35-2 of the block 35, 34-2 of the block 34, 50-2 of the shaper 50 , 40-2 of block 40, 49-2 of block 49 and 52-2 of block 52 and the fourth output 51-4 of block 51 by means of a bi-directional bus.

 The first input 53.1 of the switch 53 of the block 11 (see Fig. 11) is connected to the second input 11.2 of the block 11, the third input 11.3 of which is connected to the second input 53.2 of the switch 53 and to the first input 54.1 of the filter 54, the output of which is connected to the third input 53.3 of the switch 53 The first input 11.1 of block 11 is connected to the first input 55.1 of the analyzer 55. The output of the switch 53 is connected to the input of the BOZU 56, the output of which is connected to the first input 57.1 of the shaper 57, the fifth input 57.5 of which is connected to the first output 55-1 of the analyzer 55, and the output with the input of the node 58, the output of which is connected to the input of the BOSE 59, the output of which is connected to the sixth input 60.6 of the driver 60, the output of which is connected to the input of the node 61, the first output 61-1 of which is connected to the first input 62.1 of the switch 62, the third input 55.3 of the analyzer 55 and the second output 11-2 of block 11, the fourth input of 11.4 of which connected to the third input 60.3 of the shaper 60, the second input 57.2 of the shaper 57 and the second input 55.2 of the analyzer 55. The fifth input 11.6 of block 11 is connected to the fourth input 60.4 of the shaper 60, the sixth input 11.6 - with the third input 57.3 of the shaper 57, and the seventh input 11.7 - s fifth input 60.5 shaper 60 and h the solid input 57.4 of the shaper 57. The first input 60.1 of the shaper 60 is connected to the first input 55.1 of the analyzer 55, and the second 60.2 to the input of the BOZU 56. The first output 62-1 of the switch 62 is connected to the input of the approximator 63, the first output 63-1 of which is connected to the input node 64. The second 62-2 and third 62-3 outputs of the switch 62 are connected to the inputs of nodes 65 and 66, respectively. The first outputs of the nodes 64, 65 and 66 are combined with the fourth input 55.4 of the analyzer 55 and connected to the first output 11-1 of the block 11. The eighth input 11.8 of the block 11 is connected to the second input 54.2 of the filter 54, the fifth input 55.5 of the analyzer 55, the seventh input 60.7 of the shaper 60 and the second input 62.2 of the switch 62. The third output 11-3 of the block 11 is connected to the second outputs 55-2 of the analyzer 55, 61-2 of the node 61, 63-2 of the approximator 63, 64-2, 65-2 and 66-2 of the nodes 64, 65 and 66, respectively, via a bi-directional bus.

 The first 32.1, second 32.2, third 32.3, fourth 32.4 and fifth 32.5 inputs of the shaper 32 of block 10 are connected to the first 67.1, second 67.2, third 67.3, fourth 67.4 and fifth 67.5 inputs of the shaper 67 (see Fig. 12), respectively, the first output 67 -1 which is connected to the first input 68.1 of node 68, the output of which is the first output 32-1 of driver 32 of block 10, the sixth input 32.6 of which is connected to the second input 68.2 of node 68, and the seventh input 32.7 - with the sixth input 67.6 of driver 67 and the third input 68.3 node 68. The second output 32-2 of the shaper 32 of block 10 is connected to the second output 67-2 of the form rover 67 via a bi-directional bus

 The fifth input 36.5 of the driver 36 of block 10 is connected to the input of the BOSU 69 (see Fig. 13), the output of which is connected to the fifth input 70.5 of the driver 70, the first 70.1, second 70.2, third 70.3 and fourth 70.4 whose inputs are connected to the first 36.1, second 36.2 , the third 36.3 and fourth 36.4 inputs of the shaper 36 of block 10, respectively. The output of the driver 70 is connected to the third input 71.3 of the node 71, the output of which is connected to the input of the BOSE 72, the output of which is connected to the sixth 70.6 and the second 73.2 inputs of the driver 70 and the node 73, respectively. The first 71.1 and second 71.2 inputs of the node 71 are connected to the first 70.1 and second 70.2 inputs of the driver 70. The output of the node 73 is connected to the input of the node 74, the output of which is the output of the driver 36 of block 10, the sixth input 36.6 of which is connected to the third input 73.3 of the node 73 and the seventh the input 70.7 of the shaper 70.

 To use the invention in industry, the device according to the second embodiment can be performed, for example, as follows.

 The first input 75.1 of the block 75 of the signal processing device (see Fig. 14) is connected to the first input of the device, and the second input 75.2 to the second input of the device. The input of block 76 is connected to the third input of the device. The first output 77-1 of the block 77 is connected to the second input 78.2 of the shaper 78, the first input 78.1 of which is connected to the first output 75-1 of the block 75. The second input 77.2 of the block 77 is connected to the sixth input of the device. The output of the driver 78 is connected to the first input 79.1 of the block 79. The second output 75-2 of the block 75 is connected to the third input 77.3 of the block 77, the fourth input 78.4 of the frame shaper 78, the fifth input 79.5 of the block 79, the second input 80.2 of the block 80, the third input 81.3 of the block 81 , the eighth input 82.8 of block 82, the fourth input 83.4 of block 83, the fifth input 84.5 of switch 84, the fourth input of 85.4 of block 85, the fourth input of 86.4 of block 86, the second input of 87.2 of block 87, the fourth input of 88.4 of block 88, the fifth input of 89.5 of block 89, the third input 90.3 of block 90, sixth input 91.6 of analyzer 91, third input 92.3 of block 92, third the input 93.3 of the block 93 and the sixth input 94.6 of the shaper 94. The output of the block 76 is connected to the third input 78.3 of the shaper 78, with the fifth input 81.5 of the block 81, the first input 82.1 of the block 82, the first 83.1 and the second 88.2 inputs of the blocks 83 and 88, respectively, and the first the input 90.1 of the block 90. The fourth input of the device is connected to the first input 84.1 of the switch 84, and the fifth input is connected to the second input 84.2 of the switch 84, the first 84-1 and second 84-2 of which outputs are connected to the third 79.3 and second 79.2 inputs of the block 79, respectively as well as with the first 81.1 and second 81.2 and seventh 82.7 and sixth 82.6 block inputs 81 82, respectively. The output of block 79 is connected to the seventh input 81.7 of block 81 and to the third input 82.3 of block 82. The output of block 80 is connected to the first input 87.1 of block 87. The first output 82-1 of block 82 is connected to the second input 83.2 of block 83. The first output 81.1 of block 81 is connected with the first input 80.1 of block 80. The second output 81-2 of block 81 is connected to the third input 84.3 of switch 84, with the first input 85.1 of block 85 and with the first input 94.1 of driver 94. The third output 81-3 of block 81 is connected to the first input 88.1 of block 88 the first output 88-1 of which is connected to the fourth input 89.4 of block 89. The fourth output 81-4 of block 81 is connected to the second input 85.2 of block 85 and with the second input 91.2 of analyzer 91. The fifth output 81-6 of block 81 is connected to the first input 91.1 of analyzer 91. The sixth output 81-6 of block 81 is connected to the third input 86.3 of block 86. Seventh output 81-7 of block 81 connected to the fourth input 75.4 of block 75. The first output 83-1 of block 83 is connected to the third input 89.3 of block 89. The second output 82-2 of block 82 is connected to the first input 86.1 of block 86. The first outputs 85-1 and 86-1 of blocks 85 and 86 are connected to the first and second inputs of block 89, respectively. The second output 85-2 of block 85 is connected to the second input 86.2 of block 86, the second output 86-2 of which is connected to the third input 85.3 of block 85. The first output 87-1 of block 87 is connected to the third input 91.3 of analyzer 91. The first output 89-1 of block 89 is connected to the fifth input 82.5 of block 82, the second input 90.2 of block 90, the fourth input 91.4 of analyzer 91 and the first input 92.1 of block 92, the first output of which 92-1 is connected to the second input 93.2 of block 93 and the fifth input 91.5 of analyzer 91, the first the output 91-1 of which is connected to the fourth input 81.4 of the block 81, the second input 92.2 of the block 92 and the second input 94.2 form 94. The first output 90-1 of block 90 is connected to the fourth input 84.4 of switch 84, with the first input 93.1 of block 93 and with the third input 94.3 of driver 94. The second output 91-2 of the analyzer 91 is connected to the second input 82.2 of block 82. The first output 93 -1 block 93 is connected to the fourth input 94.4 of the driver 94. The second output 92-2 of the block 92 is connected to the fifth input 94.5 of the driver 94. The first output 94-1 of the driver 94 is connected to the output of the device and to the first input 77.1 of block 77, and the second output 94 -2 - with the fourth input 79.4 of block 79, the sixth input of 81.6 of block 81, the fourth input of 82.4 of block 82, and t by the inputs 83.3 and 88.3 of blocks 83 and 88. The fourth output 91-4 of the analyzer 91 is connected to the third input 75.3 of block 75. The processor 95 is connected to the second outputs 77-2 of block 77, 83-2 and 88-2 of blocks 83 and 88, 87 -2 blocks 87, 89-2 and 90-2 blocks 89 and 90 and 93-2 blocks 93, the third outputs 85-3 and 86-3 blocks 85 and 86, 91-3 of the analyzer 91, 92-3 blocks 92 and 94 -3 shaper 94 and the eighth output 81-8 of block 81 by means of a bi-directional bus.

 The first 6.1 and second 6.2 inputs of block 6 of the device for determining the coordinates of objects are connected to the first and second inputs of the first multiplexer (see Fig. 15). The third 6.3 and fourth 6.4 inputs of block 6 are connected to the first and second inputs of the second multiplexer. The fifth input 6.5 of block 6 is connected to the third inputs of both multiplexers, the outputs of which are outputs 6-1 and 6-2 of block 6.

 The first 13.1 and second 13.2 inputs of the switch 13 (in the first embodiment) (as well as the first 84.1 and second 84.2 inputs of the switch 84 in the second embodiment) are connected to the first inputs of the first and second multiplexers (see Fig. 16), and the third 13.3 and fourth 13.4 inputs of the switch (and the third 84.3, and the fourth 84.4 inputs - in the second embodiment) are connected to the second inputs of both multiplexers. The fifth input 13.5 of the switch 13 (or the fifth input 84.5 of the switch 84) is connected to the third inputs of both multiplexers, the outputs of which are the first and second outputs 13-1 and 13-2, respectively, of the switch 13 (or the first and second outputs 84-1 and 84-2 respectively switch 84).

 The first 31.1, second 31.2 and third 31.3 inputs of the switch 31 (see Fig. 17) of block 10 (as well as the first 81.1, second 81.2 and third 81.3 inputs of the switch of block 81 in the second embodiment of the claimed device) are connected to the first, second and third inputs of the multiplexer respectively. The first and second outputs of the multiplexer are the first 31-1 and second 31-2 outputs of the switch 31, respectively (or the first 81-1 and second 81-2 outputs of the switch 81, respectively).

 The first 53.1, second 53.2 and third 53.3 inputs of the switch 53 (see Fig. 18) of block 11 (as well as the first 82.1, second 82.2 and third 82.3 inputs of the switch 82 in the second embodiment of the inventive device) are connected to the first, second and third inputs of the multiplexer, respectively . The multiplexer output is the output of switch 53 (or switch 82).

To implement the invention according to the first embodiment, the signal processing device for determining the coordinates of objects observed in a sequence of television images may comprise, for example, a block 1 (see FIG. 8) for receiving and storing signals of the current field of the television image from a video camera of a surveillance system, received at the first input 8.1, receiving an external START / STOP signal received at the second input 8.2, receiving a disruption signal for automatic coordinate determination (AOK) received at the third input 8. 3 from block 20, and receiving a signal at input 8.4 from block 10 and generating signals for synchronizing the operation of the device, block 2, intended for receiving and storing signal codes (for example, digital) of the angular velocity of uncontrolled movement and roll of the axis of the field of view of the video camera of the surveillance system received to the input of the block from the third input of the device and their translation from the outputs of the buffer registers to the output of the block, block 3, designed to determine the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception the previous and current fields of the television image due to the action on the video camera of the monitoring system of the control signals to move its field of view, while the values of the control signal codes X _UPR [n], Y _UPR [n] are received at the first input 3.1 of the block from the shaper 23, the values of the codes the angular position of the control panel are received at the second input 3.2 of the unit from the sixth input of the device, signals of AOK, frame sync pulses (CSI) (determining the beginning of calculations in the current fields of television images (PTI), clock impu Lsovs of word-by-word synchronization of TSI video data (synchronizing the operation of the unit) are received at the third input 3.3 of the unit from unit 1 via the synchronization bus (AL); block 4, intended for generating signals of a measure of dissimilarity of television images of 2N landmarks in the current and previous fields of television images, determining the offsets of signals of television images 2N of landmarks and calculating with their use the parameters of the shift and rotation of signals of the current n-th field of the television image relative to (n -1) -th field of a television image, block 5, designed to separate received on the basis of evaluating the measure of dissimilarity of television images 2N orient ranges of terrain parameters for shifting the signals of the current field of the television image between the reception of signals from the current and previous fields of television images to the components of the controlled movement of the axis of the field of view of the video camera of the monitoring system and the uncontrolled movement of the axis of the field of view of the video camera of the monitoring system, block 6, designed to supply signals of controlled movement of the axis fields of view of the video camera of the monitoring system and signals of uncontrolled movement and roll of the field of view of the video camera of the monitoring system pressure on the device blocks directly from block 3 and block 2, respectively, or from block 5, shaper 7, designed to generate television image signals of the current nth frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between receiving signals of the (n-1) -th and (n-2) -th fields of the television image and the signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television isionic image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, block 8, designed to generate and scale the television image signals in the current analysis window using position and size signals of the current television image analysis window, block 9, designed to form M windows analysis of the background around the perimeter of the analysis window and determining the projections of the signals of the binary television image of the histogram classifier in M about background analysis buttons, block 10, designed to determine the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of the generalized binary television image of the object, block 11, designed to determine the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of measure signals dissimilarities of television images, block 12, designed to determine the current speed of movement t a levision image of an object in an inertial coordinate system using the coordinates of a television image of an object obtained on the basis of generating signals of a measure of dissimilarity of television images, the switch 13, designed for electronic switching on the device blocks of the source codes (obtained from the monitoring system) or current (obtained as a result of processing television images in the current analysis window) of the size of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, block 14, assigned to determine the reliability coefficients of the current speed of movement of the generalized binary television image of the object in the inertial coordinate system, block 15, designed to determine the reliability coefficients of the current speed of movement of the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, block 16, designed to determine the area and coordinates of the boundaries of binary television images g histogram classifier in M windows for analyzing the background from the obtained projections of the signals of the binary television image histogram classifier in M windows for analyzing the background, block 17, designed to determine the current speed of movement of the generalized binary television image of the object in an inertial coordinate system, block 18, designed to form a comprehensive assessment of the current the speed of movement of a television image of an object in an inertial coordinate system from the current speed estimation data the distribution of the television image of the object, obtained on the basis of the formation of signals of a measure of dissimilarity of the television images and the estimation of the current speed of the television image of the object from the generalized projections of the signals of the generalized binary television image of the object, taking into account the reliability coefficients of the components of the speed of the television image of the object and a priori restrictions on the speed of maneuvering the object, block 19, designed to formulate a comprehensive assessment of television the coordinates of the image of the object in the field of view of the video camera of the monitoring system by integrating the difference of the complex estimate of the current speed of the image of the object and the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system in the inertial coordinate system using the initial coordinates of the television image of the object in the field of view of the camera received from the monitoring system at the time the beginning of tracking the object, the analyzer 20, designed to analyze the current and average area of the general of a binary television image of the object, the current and average speed of moving the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the M windows for analyzing the background and generating, based on this analysis, the transition signal to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system using extrapolated coordinates of the television image of the object, block 21, designed for averaging (low-frequency iltration) a complex assessment of the current speed of moving a television image of an object and storing the values of the averaged (filtered) speed, block 22, designed to determine the predicted coordinates and speed of movement of a television image of an object in the next frame and restore the true speed of a television image of an object based on the analysis of stored values of the averaged a comprehensive assessment of the speed of movement of a television image of an object when switching to formations the control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates, the shaper 23, designed to generate control signals for the movement of the axis of the field of view of the video camera of the surveillance system and signals of position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view video cameras of a surveillance system or extrapolated coordinates and speed of a television image moving about the object and the processor 24, designed for the computational processing of local data transmitted over a common bi-directional bus (including virtually implemented parallel branches of the algorithm) that do not require stream processing of large data arrays.

Block 4 contains, for example, (see Fig. 9) a selector 25 for selecting signals of television images of three pairs of terrain landmarks from signals of the current television image field, a calculator 26. for generating signals of a measure of dissimilarity of television images of the first pair of terrain landmarks from signals television images of the first pair of landmarks of the current field of the television image and signals of television images of the corresponding first pair of landmarks in the previous present television picture field and determining their images via displacement
the first pair of terrain landmarks during the time between receiving signals of the current and previous fields of television images, the calculator 27, designed to generate signals of dissimilarity measures of television images of the second pair of terrain landmarks from the signals of the television images of the second pair of landmarks of the current field of the television image and the television images of the corresponding second pair landmarks in the previous field of the television image and determination using them offsets and images of the second pair of landmarks during the time between the reception of signals of the current and previous fields of television images, the calculator 28, designed to generate signals of the measure of dissimilarity of the television images of the third pair of landmarks from the signals of the television images of the third pair of landmarks of the current field of the television image and the signals of the television images corresponding to the third pairs of landmarks in the previous field of the television image and definitions with their help of the displacements of the images of the third pair of landmarks of the terrain between the reception of signals of the current and previous fields of television images, the calculator 29, designed to determine using the offsets of the coordinates of the signals of the television images of three pairs of landmarks, the parameters of the shift and rotation of the signals of the current field of the television image relative to the previous field of the television image .

 Moreover, in this example implementation of the device to simplify the consideration, the number N = 3 of the used pairs of landmarks is adopted.

среднее значение сигналов телевизионного изображения в окне объекта,

среднее значение сигналов телевизионного изображения в окне фона,
σ_ф - среднеквадратическое отклонение сигналов телевизионного изображения в окне фона от

формирователь 36, предназначенный для формирования сигналов первичного бинарного телевизионного изображения детектора изменений фона посредством образования разностного телевизионного изображения детектора изменений фона вычитанием из сигналов масштабированного изображения в текущем окне анализа сигналов телевизионного изображения эталонного фона и сравнения полученных сигналов разностного телевизионного изображения детектора изменений фона с порогом, фильтр 37, предназначенный для низкочастотной пространственной фильтрации сигналов первичного бинарного телевизионного изображения детектора изменений фона, формирователь 38, предназначенный для формирования сигналов вторичного бинарного телевизионного изображения детектора изменений фона, формирователь 39, предназначенный для формирования горизонтальных и вертикальных проекций сигналов вторичного бинарного телевизионного изображения детектора изменений фона, блок 40, предназначенный для определения коэффициентов достоверности сигналов вторичных бинарных телевизионных изображений детектора изменений фона, второе БОЗУ 41, предназначенное для запоминания сигналов масштабированного телевизионного изображения в текущем окне анализа, узел 42, предназначенный для приведения к текущему масштабу в окне анализа сигналов масштабированного телевизионного изображения и смещения их на величину перемещения оси поля зрения видеокамеры системы наблюдения, БОЗУ 43, предназначенное для запоминания приведенных к текущему масштабу в окне анализа и смещенных на величину перемещения оси поля зрения видеокамеры системы наблюдения сигналов масштабированного телевизионного изображения, узел 44, предназначенный для формирования сигналов разностного телевизионного изображения селектора движущихся объектов, формирователь 45, предназначенный для формирования сигналов первичного бинарного телевизионного изображения селектора движущихся объектов, фильтр 46, предназначенный для низкочастотной пространственной фильтрации сигналов первичного бинарного телевизионного изображения селектора движущихся объектов, формирователь 47, предназначенный для формирования сигналов вторичного бинарного телевизионного изображения селектора движущихся объектов, формирователь 48, предназначенный для формирования горизонтальных и вертикальных проекций сигналов вторичных бинарных телевизионных изображений селектора движущихся объектов, блок 49, предназначенный для определения коэффициентов достоверности сигналов вторичных бинарных телевизионных изображений селектора движущихся объектов, формирователь 50, предназначенный для формирования обобщенных горизонтальных и вертикальных проекций сигналов обобщенного бинарного телевизионного изображения объекта, блок 51, предназначенный для определения координат и размеров, текущей и усредненной площади обобщенного бинарного телевизионного изображения объекта в окне анализа, и блок 52, предназначенный для анализа условий срыва автоматического определения координат телевизионного изображения объекта и формирования признака Рr_-АОК - прекращения автоматического определения координат объекта.Block 10 contains, for example, (see FIG. 10) a first buffer random access memory (BOS) 30 for storing the signals of the scaled television image in the current analysis window, a switch 31 for electronic switching to the shaper 32 of the signals of the scaled television image in the current analysis window directly from block 8 (see Fig. 8) or after buffering in BOZU 30, shaper 32, intended for generating signals of a binary television image of histograms classifier in the current analysis window, shaper 33, designed to generate horizontal and vertical projections of the binary television image signals of the histogram classifier, block 34, used to determine the reliability coefficients of binary television image signals of the histogram classifier, block 35, used to determine the object / background Q and the minimum value _fmin σ (n) - rms signals television picture background in the windows M NALYSIS background,

the average value of the television image signals in the object window,

the average value of the television image signals in the background window,
σ _f - the standard deviation of the signals of the television image in the background window from

a shaper 36 intended for generating signals of a primary binary television image of a background change detector by generating a differential television image of a background change detector by subtracting a background image from the signals of a scaled image in the current window for analyzing signals of a television image and comparing the received signals of a difference television image of a background change detector with a threshold, a filter 37, designed for low-pass spatial filtering ignals of the primary binary television image of the background change detector, shaper 38 for generating signals of the secondary binary television image of the background change detector, shaper 39 for forming horizontal and vertical projections of signals of the secondary binary television image of the background change detector, block 40 for determining coefficients the reliability of the signals of the secondary binary television images of the change detector background, the second BOSU 41, designed to store the signals of the scaled television image in the current analysis window, node 42, designed to bring to the current scale in the analysis window of the signals of the scaled television image and shift them by the amount of movement of the axis of the field of view of the video camera of the surveillance system, BOSU 43, designed to store reduced to the current scale in the analysis window and offset by the magnitude of the displacement axis of the field of view of the video camera of the signal observation system of scales the actual television image, the node 44, intended for generating signals of the differential television image of the selector of moving objects, the shaper 45, for generating signals of the primary binary television image of the selector of moving objects, a filter 46, designed for low-frequency spatial filtering of the signals of the primary binary television image of the selector of moving objects, driver 47 for generating signals of a secondary binary of the television image of the moving object selector, shaper 48, designed to generate horizontal and vertical projections of the signals of the secondary binary television images of the moving object selector, block 49, designed to determine the reliability coefficients of the signals of the secondary binary television images of the selector of moving objects, shaper 50, designed to generate generalized horizontal and vertical projections of the signals of a generalized binary television the image of the object, block 51, designed to determine the coordinates and sizes, the current and average area of the generalized binary television image of the object in the analysis window, and block 52, designed to analyze the conditions for the breakdown of the automatic determination of the coordinates of the television image of the object and the formation of the sign of Pr- _AOK - termination automatic determination of the coordinates of the object.

 Block 11 contains, for example, (see Fig. 11) a switch 53 intended for electronic switching to the device blocks of signals of a scaled television image in the current analysis window directly from block 8 (see Fig. 8) or passed through a non-linear high-pass filter 54 , an analyzer 55, designed to compare the parameters of the signals of the measure of dissimilarity and the parameters of the trajectories of the television image of the object obtained using static and dynamic reference images of the object or only with static reference television image of the object and generate the update signal of the static reference television image of the object, BOZU 56, intended for storing signals of the scaled television image in the current analysis window, shaper 57, designed to generate signals of static and dynamic reference images of the object or signals of only static reference television image object, node 58, designed to bring the signals of static and of the physical reference images of the object or signals of only the static reference television image of the object to the current scale, BOSE 59, designed to bring the signals of static and dynamic reference images of the object or signals of only the static reference television image of the object to the current scale, shaper 60, intended to bring the signals of static and dynamic reference images of an object or signals of only a static reference television image object to the current scale, node 61, designed to determine the minimum values of the signals of the measure of dissimilarity of television images along the rows and along the columns of the two-dimensional region of the search for displacements of the television image of the object, the formation and determination of the type of sequences of the minimum values of signals of the measure of dissimilarity of images along the rows and along the columns of the two-dimensional region the search for the displacement of the television image of the object, the switch 62, designed for switching data sequences of minimum the values of the signals of the measure of dissimilarity of television images to the input of one of the three nodes 64, 65 or 66, the approximator 63, designed for analytical approximation of the sequence of the minimum values of the signals of the measure of dissimilarity of television images by a fourth degree polynomial, the node 64, designed to determine the coordinates of the television image of the object in the window analysis of the minimum position of the approximating polynomial of the fourth degree of the data sequence of the minimum values of the signals of the measure of dissimilarity television images, node 65, designed to determine the coordinates of the television image of the object in the analysis window by shifting the boundary of the region of rapid growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window, node 66, designed to determine the coordinates of the television image of the object in the analysis window by the coordinates of the center of the window analysis.

 Shaper 32 (see Fig. 12) of block 10 (see Fig. 8) contains, for example, shaper 67, designed to generate normalized histograms of the distribution of brightness of television images of the object and background, and node 68 (histogram classifier), designed to classify elements a television image in the current window for analyzing the elements of an object and background based on a comparison of normalized histograms of the distribution of brightness of television images of an object and background with the formation of a binary television image of a histogram Amm classifier.

Shaper 36 (see Fig. 13) of block 10 (see Fig. 8) contains, for example, BOSU 69, intended for storing signals of a scaled television image in the current analysis window, shaper 70, intended for generating signals of a television image of a reference background, node 71, designed to bring the television image signals of the reference background obtained in the previous n-1 frame to the current scale and shift, taking into account the movement of the center of the analysis window in the inertial coordinate system for the last frame, B RAM 72, intended for storing the television image signals of the reference background obtained in the previous n-1 frame, reduced to the current scale and offset by the amount of displacement of the center of the analysis window in the inertial coordinate system for the last frame, node 73, designed for generating differential television image signals the detector for changing the background by subtracting from the signals of the scaled image in the current window of the analysis of the signals of the television image of the reference background, and the node 74, intended for I determine the threshold _DIF Threshold (ix, jy) of binarization of the background change detector as a value proportional to the local scattering parameter of the difference television image of the background change detector in the vicinity of the point with coordinates ix, jy and the formation of the primary binary image L _{1 BIN DIF} (ix, jy) background change detector in accordance with the rule:
L _{1 BIN DIF} (ix, jy) = 1 if | L _{p DIF} (ix, jy) | ≥POROG _DIF (ix, jy),
or
L _{1 bin differential} (ix, jy) = 0 if | L _{p DIF} (ix, jy) | < _DIF THRESHOLD (ix, jy).
For the implementation of the invention according to the second embodiment, the signal processing device may comprise, for example, a block 75 (see FIG. 14) for receiving and storing signals of the current field of a television image from a video camera of a surveillance system and generating synchronization signals of the device, block 76 for receiving and storing signals of uncontrolled movement and roll of the field of view of the video camera of the surveillance system, block 77, designed to determine the controlled movement of the axis of the field of view of the video camera with observation systems for the time between the reception of signals of the previous and current fields of the television image, due to the action on the video camera of the monitoring system of the motion control signals of its field of view, shaper 78, designed to generate television image signals of the current nth frame from the television image signals of the previous frame, taking into account the controlled moving the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the (n-1) -th and (n-2) -th fields of the television image signals and the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, block 79, designed to generate and scale the signals of the television image in the current analysis window using position and size signals the current window of the analysis of the television image, block 80, designed to form M windows analysis of the background according to the window size of the analysis and determination of the projections of the signals of the binary television image of the histogram classifier in M background analysis windows, block 81, designed to determine the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, block 82, designed to determine coordinates of the television image of the object relative to the center of the current analysis window based on the the signals of the measure of dissimilarity of television images, block 83, designed to determine the current speed of movement of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, switch 84, designed for electronic switching to device blocks source codes (obtained from the surveillance system) or current (received as a result of processing television units in the current analysis window) of the size of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, block 85, designed to determine the reliability coefficients of the current speed of movement of the generalized binary television image of the object in the inertial coordinate system, block 86, designed to determine the reliability coefficients of the current speed television image of an object in an inertial coordinate system, obtained on the basis of the formation of signals of a measure of dissimilarity of bodies of visual images, block 87, intended for determining the area and coordinates of the boundaries of binary television images of a histogram classifier in M background analysis windows from the obtained projections of binary television image signals of a histogram classifier in M background analysis windows, block 88, for determining the current speed of movement of a generalized binary television images of an object in an inertial coordinate system, block 89, designed to form a comprehensive assessment of the current speed the growth of the television image of the object in the inertial coordinate system from the data of the current television image velocity of the object obtained from the signal generation based on the signal dissimilarity of the television images and the current speed of the television image of the object from the generalized signal projections of the generalized binary television image of the object, taking into account the reliability factors components of the speed of moving a television image of an object and apr ioor restrictions of the object’s maneuvering speed, block 90, intended for the formation of a complex estimate of the television coordinates of the image of the object in the field of view of the video camera of the monitoring system by integrating the difference of the complex estimate of the current speed of the image of the object and the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system in the inertial coordinate system using the initial coordinates of the television image of the object in the field of view of the camera, obtained from systems observations at the moment of starting tracking the object, analyzer 91, designed to analyze the current and average area of the generalized binary television image of the object, the current and average speed of movement of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in M windows for analysis of the background and generation of the basis of this analysis of the transition signal to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system using by extrapolated coordinates of the television image of the object, block 92, designed to average (low-pass filtering) a comprehensive estimate of the current speed of the television image of the object and remember the values of the averaged (filtered) speed, block 93, designed to determine the predicted coordinates and speed of the television image of the object in the following frame and restore the true speed of the television image of the object based on the analysis of memorized values of the averaged complex estimate of the speed of moving the television image of the object during the transition to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates, the shaper 94 designed to generate the control signals for the movement of the axis of the field of view of the video camera of the surveillance system and the signals of the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in For the view of the video camera of the monitoring system or the extrapolated coordinates and the speed of the television image of the object, and the processor 95, designed for the computational processing of local data transmitted via a common bi-directional bus (including virtually realized parallel branches of the algorithm) that do not require stream processing of large data arrays.

 Block 81 can be implemented in the same way as block 10 according to the circuit shown in FIG. 10.

 Block 82 can be implemented according to the circuit shown in Fig. 11, i.e. same as block 11.

An example of the implementation of the first proposed embodiment of the device can be given for the execution case, in which:
- the functions of streaming computational processing and control logic are implemented as functions of programmable logic integrated circuits (FPGAs) on microcircuits, for example, of the type EPF10K130E ... from Altera Co. (USA);
- the functions of the computational processing of local data (not requiring stream processing of large data arrays) are implemented in a common processor on a chip, for example, such as TMS320C80 ... from Texas Instruments Inc. (USA) used by the corresponding units of the device in the separation of time on a common bi-directional bus (OS) (including virtually when implementing the corresponding parallel branches of the algorithm);
- small arrays of processed data (variable parameters) are stored in the internal RAM of the FPGA and the processor 24;
- the constant parameters of stream processing and control are stored in internal registers and RAM of the FPGA (loaded together with loading the configuration of the FPGA when the power is turned on from boot ROMs sold on microcircuits, for example, type EPCLI20) by Altera Co. (USA);
- large arrays of processed data are stored in autonomous (external to the FPGA) RAM, implemented on standard microcircuits, for example, type IDT71V424 ... from Integrated Device Technology (USA);
- programs (subprograms) and processor processing parameters are stored in the external RAM of the processor sold on microcircuits, for example, of the MT48LC1M16 type. . . Micron technology Inc. (USA) (loaded by the processor when the power is turned on from a bootable FLASH ROM sold on a chip, for example, type AM29LV017В ...) Advanced Micro Devices Inc. (USA).

Block 1 can be implemented, for example, on the basis of autonomous RAM and the following FPGA functions:
- a communication channel for external video data (for example, a digital video signal) received at the first input 1.1 of block 1 (see Fig. 8.1), with an input of autonomous RAM (recording channel);
- a communication channel of autonomous RAM by video data with the first output 1-1 of block 1 (read channel);
- addressing the recording channel by receiving in standalone RAM video data of the current field of the television image (PTI));
- addressing of the read channel for the issuance of autonomous RAM video data of the previous IPT;
- extracting from the digital full video signal video synchronization signals: frame sync pulses (CSI), horizontal sync pulses (SSI), word-by-word synchronization clocks of video data (TSI) for use in other functions of unit 1 and functions of other device blocks - are transmitted as an integral part of the synchronization bus ) to the second output of 1-2 blocks 1;
- generating a signal of the automatic coordinate determination mode (AOK: AOK = 1 - beginning of a mode, AOK = 0 - termination) based on an external START / STOP signal (see steps 108/110 of FIG. 19) received at the second input 1.2 of block 1 and AOK failure signal (action 143 in Fig. 19), supplied to the third input 1.3 of block 1 from the driver 50 (see Fig. 10.2), for use in other functions of block 1 and the functions of other blocks of the device - it is transmitted as an integral part of the loop to the second output 1-2 of block 1;
- generation of the code n of the PTI number (with AOK = 0-n = 0, with AOK = 1 - the value of n is equal to the current number of PTI from the beginning of the AOK) - it is transmitted as an integral part of the loop to the second output 1-2 of block 1.

 Block 2 can be implemented, for example, as an FPGA function for buffering (in internal FPGA registers) the signal codes (for example, digital) of the angular velocity of the axis of the field of view of the video camera of the surveillance system received at the input of block 2 from the third input of the device and broadcasting them from the outputs of the buffer registers to the output of block 2.

Block 3 can be implemented, for example, by the following FPGA functions:
- call the subroutine for the implementation of expressions (39) in the processor 24 by issuing the corresponding interrupt to the shared bus to the second output 3-2 of block 3, provided AOK = 1, otherwise:
- assigning to the displacements dx and dy the values of the codes calculated (action 99 of the algorithm in Fig. 19) from the values of the codes of the angular position of the control panel received at the second input 3.2 of block 3 (action 98 of the algorithm in Fig. 19);
wherein:
- the values of the control signal codes X _UPR [n], Y _UPR [n] are received from the first input 3.1 of block 3 from the shaper 23 and are buffered in the internal FPGA registers programmatically accessible to the processor 24 via the OS connected to the second output 3-2 of block 3, in read mode;
- signals of AOK, CSI (determining the beginning of calculations in the current PTI), TSI (synchronizing the operation of the block) are received by the loop from the third input 3.3 of block 3 from block 1;
- the routine for the implementation of expressions (39) and the values of the impulse characteristics hx [i], hy [i] are stored in the RAM of the processor 24; the resulting values dx [n] and dy [n] are buffered in the internal FPGA registers programmatically accessible to the processor 24 via the OS in write mode, and issued (from the outputs of the buffer registers) to the first output 3-1 of block 3, etc.

 The entire operation of the first claimed embodiment of the device, implemented for conditions when there are low-frequency uncontrolled movements of the field of view of the video camera of the surveillance system (which lead to displacements and rotations of all elements of the field of the television image as a whole), as well as the possible absence of stabilization error sensors in the monitoring system the field of view of the video camera of the surveillance system is illustrated by the flowchart of the algorithm of operation in Fig.19.

 The operation of the second claimed embodiment of the device, implemented for conditions when there are wideband uncontrolled movements of the field of view of the video camera of the surveillance system ("jitter" of the field of view of the video camera of the monitoring system) with greater dynamics of the object, is illustrated by the flowchart of FIG. 20.

Claims (54)

 1. A method of processing signals to determine the coordinates of objects observed in a sequence of television images, which consists in receiving and storing the signals of the current field of the television image, selecting from the signals of the current field of the television image the signals of the television images 2N landmarks, and N = 3, 4 , 5, . . . the signals of the dissimilarity of the television images 2N of the terrain landmarks are generated from the signals of the television images 2N of the terrain landmarks of the current field of the television image and the signals of the television images of the corresponding 2N landmarks in the previous field of the television image, and the bias of the television images 2N of the terrain landmarks is determined using them for the time between signal reception previous and current fields of television images are determined using offsets of TV signals 2N landmarks, the shift and rotation parameters of the signals of the current field of the television image during the time between receiving signals from the previous and current fields of television images, characterized in that the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals from the previous and current fields of the television image, due to the action on the camcorder of the monitoring system for signals to control the movement of its field of view, determine before receiving signals The fraction of the television image, the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system is determined from the data of the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals from the previous and current fields of the television image, and the signals of the current field of the television image are received and stored and roll the field of view of the video camera of the surveillance system and use them to generate signals, we predict coordinates of television images of 2N landmarks in the current field of the television image, divide the parameters of the shift of the signals of the current field of the television image obtained between the signals of the current and previous fields of television images from the components of the controlled movement of the axis of the field of view video cameras of the surveillance system and uncontrolled movement of the axis of the field of view video cameras of the surveillance system, forms signals of the television image of the current frame from the signals of the television image of the previous frame taking into account the controlled movement of the axis of the field of view during the time between receiving signals of the previous and current fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, form and scale the signal s of the television image in the current analysis window using the position and size signals of the current television image analysis window obtained by processing the television image in the previous frame or using the extrapolated coordinates and speed of the television image of the object, with the initial conditions of the position signals and sizes of the current television analysis window images formed at the moment of starting tracking an object based on an external START / STOP signal are stored the signals of the scaled television image in the current analysis window generate the signals of the differential television image of the moving object selector by subtracting the signals of the scaled television image from the signals of the scaled television image in the current analysis window of the signals of the scaled television image previously stored and reduced to the current scale of the television image in the analysis window and shifted by the axis displacement fields of view of the video camera of the surveillance system, form the signals of the primary binary the evisional image of the moving object selector from the differential television image signal of the moving object selector, the signals of the secondary binary television image of the moving object selector are generated from the signals of the primary binary television image of the moving object selector that have passed low-pass filtering, simultaneously with the storage of the signals of the scaled television image in the current analysis window, by the formation differential television image signals I selectors of moving objects and the formation of signals of primary and secondary binary television images of the selector of moving objects from the signals of the scaled television image in the current analysis window and taking into account the signals of the controlled movement of the axis of the field of view of the video camera of the surveillance system form the signals of the primary binary television image of the detector for changing the background and the signals of the binary television histogram classifier images in the current analysis window, from signals primarily of a binary television image of a detector of background changes that have passed low-pass filtering, the signals of a secondary binary television image of a detector of background changes are generated, horizontal and vertical projections of signals of a secondary binary television image of a moving object selector and a detector of background changes, and horizontal and vertical projections of signals of a binary television image of a histogram classifier are generated , determine the reliability coefficients of the signals of the secondary x binary television images of the background change detector and the moving object selector, as well as signals of the binary television image of the histogram classifier, form the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object from the horizontal and vertical projections of the signals of the secondary binary television images of the selector of moving objects and the background change detector as well as from horizontal and vertical projections of binary signals the histogram image of the histogram classifier based on their joint processing using the reliability coefficients of the signals of binary television images of the histogram classifier, the signals of the secondary binary television images of the background change detector and the moving object selector, determine the horizontal and vertical borders, as well as the dimensions of the television image of the object by the cutoff levels on the left and right, above and below the set percentage of the area of the generalized horizontal and vert the projection of the signals of the generalized binary television image of the object, determine the current and average areas of the generalized binary television image of the object located inside the formed boundaries of the television image of the object, determine the current coordinates of the generalized binary television image of the object, using the generalized horizontal and vertical projections of the generalized binary television image of the object, determine generalized binary current speed of a television image of an object in an inertial coordinate system, the coefficient of reliability of the current speed of a generalized binary television image of an object in an inertial coordinate system is determined, simultaneously with the formation of signals of primary and secondary binary television images of a moving object selector and a background change detector, a binary television image of a histogram classifier, and also generalized horizontal and vertical projections of signals the binary television image of the object is determined by the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images as a result of non-linear high-frequency filtering of the signals of the scaled television image in the current analysis window, if the averaged area of the generalized binary television image of the object exceeds the threshold value , storing received signals, forming I of the signals of the static reference television image of the object or the signals of the static and dynamic reference television images of the object, bringing the signals of the static reference television image of the object or signals of the static and dynamic reference television images of the object to the current scale, the formation and storage of signals of the measure of dissimilarity between the signals of the television images after non-linear high-frequency filtering signals of a scaled television image images in the current analysis window and the signals of static or static and dynamic reference television images of the object in the two-dimensional region for searching for displacements of the television image of the object, determining the minimum values of signals of the measure of dissimilarity of television images along the rows and along the columns of the two-dimensional region for searching for displacements of the television image of the object, forming sequences of minimum values signals of dissimilarity measures of television images along rows and along columns two-dimensionally of the nth search area for the television image of the object, determining the corresponding coordinate of the television image of the object in the analysis window depending on the type of the sequence of the minimum values of the signals of the measure of dissimilarity of the television images for this coordinate, namely, by analytical approximation of the sequence of the minimum values of the signals of the measure of dissimilarity of the television images by a fourth degree polynomial and determining the coordinates of the television image of the object as a position the approximating polynomial minimum subject to assigning the sequence of minimum values of the signals of the measure of the dissimilarity of television images to the type of sequences with two boundaries of the areas of rapid growth of the values of the signals of the measure of the dissimilarity of television images near the position of its minimum, or by determining the offset of the boundary of the area of rapid growth of the values of the signals of the measure of dissimilarity of television images relative to the center the analysis window and the formation of the object coordinates in the analysis window as a value, proportional to the obtained shift of the boundary of the region of rapid growth of signal values of the measure of dissimilarity of television images, provided that the sequence of the minimum signal values of the measure of dissimilarity of television images is classified as sequences with a flat neighborhood of the minimum position and the presence of one region of rapid growth of signal values of the measure of dissimilarity of television images equal to the coordinate of the center of the analysis window, provided that the sequence the minimum values of the signals of the measure of dissimilarity of television images to the type of sequences with a flat neighborhood of the position of the minimum values of signals of the measure of dissimilarity of television images in the entire search area of the displacements of the television image of the object, determine the current speed of movement of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object obtained based on signal generation measures of dissimilarity of television images , determine the reliability coefficient of the current speed of moving the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, form a comprehensive assessment of the current speed of moving the television image of the object in the inertial coordinate system from the data of the estimation of the current speed of moving the television image of the object obtained on the basis of signal formation measures of dissimilarity of television images and the assessment of the current the speed of moving the object over the generalized projections of the signals of the generalized binary television image of the object, taking into account the reliability coefficients of the components of the speed of moving the television image of the object and the a priori restrictions on the speed of maneuvering of the object, average the signals of the integrated estimation of the speed of the television image of the object in the inertial coordinate system and store them, determine the coordinates a television image of an object in the field of view of a video camera system is observed by integrating the difference of the complex estimate of the current speed of moving the television image of the object in the inertial coordinate system and the speed of the controlled movement of the axis of the field of view of the video camera of the surveillance system in the inertial coordinate system with the initial coordinate conditions and the speed of movement of the television image of the object, formed at the start of tracking the object based on the external the START / STOP signal, simultaneously with the formation of the signals of the primary and secondary binary television images The moving object selector and the background change detector, the binary television image of the histogram classifier, the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object form the background analysis windows along the perimeter of the analysis window and determine the signal projections of the binary television image of the histogram classifier in the background analysis windows, determine the areas of the analysis and coordinates of the boundaries of binary television images of the histogram classifier in the window analyzing the background from the obtained projections of the signals of the binary television image of the histogram classifier in the background analysis windows, generate control signals for moving the axis of the field of view of the video camera of the monitoring system using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window or using the extrapolated coordinates and speed of the television image of the object depending on the results of the analysis of the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, and the extrapolated speed of the television image of the object is formed based on the analysis of the stored values of the average complex estimates of the speed of movement of the television image of the object, form signals of the position and size of the current analysis window of the television image for the next frame, using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system, obtained by processing the television image in the current analysis window, or using the extrapolated coordinates and speed of the television image of the object with the initial conditions, formed at the moment the tracking of the object starts based on the external START / STOP signal. 2. The method according to p. 1, characterized in that the controlled movement dx [n] and dy [n] of the axis of the field of view of the camera horizontally and vertically, respectively, is determined by calculating the convolution of control signals X _ypp [i], Y _ypp [i] by moving the field view of a video camera of a surveillance system with impulse characteristics hx [i] and hy [i] of its drives

where n is the number of the current half frame;
i is the element number of the impulse response;
K is the length of the impulse response. 3. The method according to p. 1, characterized in that the formation of signals L _KADR (ix, jy) of the television image of the current frame when interlaced from signals L _P-KADR (i, p, npk) of the current half-frame of the television image and television image L _{- 1KADR} (ix, jy) of the previous frame is carried out by predicting the signals L of the _FRAME (ix, jy) of the television image of the current frame by shifting the television image L _-1KADR (ix, jy) of the previous frame by the amount of controlled movement dx, dy of the field of view of the video camera of the surveillance system in the mountains isontal and vertical, respectively, during the time between half-frames of a television image
L _FRAME (ix, jy) = L _{-1 FRAME} (ix + dx, jy + dy),
and replacing the points of the television image of the current frame with the points of the television image of the current half frame with compensation for the current uncontrolled displacements r _x , r _y and roll φ of the field of view of the video camera of the surveillance system
L _KADR [ix (i, p, npk), jy (i, p, npk)] = L _P-KADR (i, p, npk),
where i is the number of the element in the line of the television image of the current half frame, i = 1,. . . Nk;
ix is the number of the element in the line of the television image of the current frame,
ix = 1,. . . , Nk;
p is the line number in the half frame,

jy is the line number in the frame, jy = 1,. . . , Mk;
Nк is the number of elements of a television image in a row;
Mk - the number of lines in the frame of the television image;
npk - index of the current half-frame:
npk = 1 - in odd half frames,
npk = 0 - in even half frames,

4. The method according to claim 1, characterized in that the signals L _{1BIN DIF} (ix, jy) of the primary binary television image of the background change detector are formed by converting the signals L _{n-1 eff} (ix, jy) of the television image of the reference background obtained in the previous n -1 frame, to the current scale by the formation of signals L _{P DIF} (ix, jy) of the differential television image of the background change detector by subtracting from the signals L _{n OA} (ix, jy) the scaled image in the current signal analysis window L _{n-1 eff} (ix, jy ) television image of the reference background of the previous frame a shift that takes into account the movement of Vx, Vy center of the analysis window in the inertial frame of the last coordinates:
L _rDIF (ix, jy) = L _{n OA} (ix, jy) -L _{n-1 eff} (ix + Vx, jy + Vy),
determination threshold binarization THRESHOLD _DIF (ix, ju) detector background changes as the value proportional to the local parameter of the scattering difference television image signal values L _rDIF (ix, ju) detector background changes in the neighborhood of the point with coordinates ix, jy, assigning values of the primary binary television picture L _{1BIN DIF} (ix, jу) background change detector

or

moreover, the signals L _{n eff} (ix, jy) of the television image of the reference background are formed by dividing the signals of the scaled television image in the current analysis window into three types of television image signals: the television image signals in the object window — OO _DIF , the television image signals in the background window — OP _DIF , the signals of the television image in the window - "New background" - NF, where as the signals of the television image in the window of the object determine the signals of the television image in the rectangle located in the center of the current analysis window and which mainly includes elements of the television image of the object, the image elements at the external borders of the current analysis window are determined as signals of the television image of the "New background" window, on which the surveillance system video cameras appear due to the movement of the object and the movement of the axis of the field of view new elements of the television image of the background, as the signals of the television image of the background window, determine all the remaining elements of the television image of the window anal from memorizing in the New Background window the signals L _nOA (ix, jy) from the current analysis window of the scaled television image of the current nth frame:
L _{n eff} (ix, jy) = L _nOA (ix, jy), ix, jy∈NF,
where L _nOA (ix, jy) are the values of the brightness signals of the element of the scaled television image in the analysis window with the coordinates ix, jy,
averaging in the background window the signals of the scaled television image L _nOA (ix, jy) from the current analysis window with a constant Wof and taking into account the shift of the analysis window in the inertial coordinate system for the last frame:
L _{n eff} (ix, jy) = (1-Wof) * L _{n-1 eff} (ix + Vx, jy + Vy) + Wo * L _{n OA} (ix, jy) ix, iy∈OF _DIF ,
where Vx, Vy - moving the center of the analysis window for the last frame horizontally and vertically, respectively, in an inertial coordinate system,
overwriting in the window of the object of the signals of the television image the reference background of the previous frame with a shift that takes into account the movement of the center of the analysis window for the last frame:
L _{n eff} (ix, jy) = L _{n-1 eff} (ix + Vx, jy + Vy), ix, iy∈OO _DIF . 5. The method according to p. 4, characterized in that the low-pass filtering of the signals of the primary binary television image L _{1-BIN DIF} (ix, jy) detector changes in the background is carried out using two-dimensional convolution

where NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
h _HF [di, dj] - impulse response of a low-pass filter. 6. The method according to p. 1 characterized in that the low-pass filtering of the signals of the primary binary television image L _{1 BIN SDO} (ix, jy) of the selector of moving objects is carried out using two-dimensional convolution

where NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
h _LF [di, dj] impulse response of the low-pass filter. 7. The method according to p. 5, characterized in that the signals L _{2 BIN DIF} (ix, jy) of the secondary binary television image of the detector for changing the background are formed from the signals of the low-pass filter S_ fil _DIF (ix, jy) in accordance with the rule:
L _{2 BIN} (ix, jy) = 1, if S_ fil _DIF (ix, jу)> Рrogrog _DIF 1 and L 1 _{BIN DIF} (ix, jy) = 1
or S_ fil _DIF (ix, jy)> Rogrog _DIF 0 and L _{1 BIN DIF} (ix, jy) = 0,
otherwise L _{2 BIN} (ix, jy) = 0,
where Porog _DIF 1, Porog _DIF 0 are the decision thresholds for the unit and zero elements of the primary binary television image of the background change detector, respectively. 8. The method according to p. 6, characterized in that the signals of the secondary binary television image L _{2 BIN SDO} (ix, jy) of the selector of moving objects are formed from signals S_ fil _SDO (ix, jy) of a low-pass filter in accordance with the rule:
L _{2BIN LMS} (ix, ju) = 1 if S_ fil _LMS (ix, jy)> Porog _DLS 1 and L _{1BIN LMS} (ix, jy) = 1
or S_ fil _LMS (ix, jy)> Porog _DLS 0 and L _{1BIN LMS} (ix, jy) = 0,
otherwise _2BIN L _SDS (ix, jy) = 0,
where Rorog _SDO 1, Porog _SDO 0 - values of decision thresholds for single and zero elements of the primary binary television image of the selector of moving objects, respectively. 9. The method according to p. 8, characterized in that the horizontal and vertical projections G _{for SDO} (ix), V _{for SDO} (jy) signals of the secondary binary television image L _{2BIN SDO} (ix, jy) of the moving object selector is determined in accordance with

for ix = 1,. . . , Nwin,

for jу = 1,. . . , Mwin,
where Nwin and Mwin are the sizes of the secondary binary television image of the selector of moving objects horizontally and vertically, respectively. 10. The method of claim. 7, characterized in that the horizontal and vertical projections G _{prosp DIF} (ix), V _{ave DIF} (jy) of the secondary binary signal television image L _{2BIN DIF} (ix, jy) detector background changes determined in accordance with

for ix = 1,. . . , Nwin,

for jу = 1,. . . , Mwin,
where Nwin and Mwin are the sizes of the secondary binary television image of the background and horizontal changes detector, respectively. 11. The method according to p. 1 and p. 4, characterized in that the horizontal and vertical projections of _{Gp GK} (ix), V _prgk (jу), signals of a binary television image L _{BIN GK} (ikh, jy) of the histogram classifier are determined in accordance from

for ix = 1,. . . , Nwin,

for jу = 1,. . . , Mwin,
where Nwin and Мwin are the dimensions of the binary television image of the histogram classifier horizontally and vertically, respectively. 12. The method according to p. 1, characterized in that the confidence coefficients W _DIF , W _SDO , W _{GK are} defined as the product of the input functions of the initial conditions and the normalized average densities of binary television images of the background change detector, moving object selector and histogram classifier, respectively

moreover, the averaged densities

binary television images of the background change detector, moving object selector, and histogram classifier are obtained by limiting the minimum and maximum values and subsequent averaging of the first densities V _DIF (n), V _SDO (n), V _GK (n) of the corresponding binary television images
Where

S _ODIF (n), S _OSDO (n), S _OGK (n) - current areas of binary television images of the background change detector, moving object selector and histogram classifier, respectively, inside the boundaries of the television image of the object;
S _OO (n) is the current area of the region inside the boundaries of the television image of the object;
F _{NU_DIF} (n), F _{NU_SDO} (n), F _{NU_GK} (n) - the input function of the initial conditions for the detector of changes in the background, the selector of moving objects and the histogram classifier;
n is the number of the current frame. 13. The method according to PP. 9 10 11 and 12, characterized in that the generalized horizontal and vertical projections _{Gp O} (ix, n), _{Vp O} (jy, n) of the signals of the generalized binary television image of the object are formed by weighted summation of the projections of the binary television images of the background change detector, selector moving objects and bar graph classifier
G _pro (ix, n) = W _DIF (n) * G _prodif (ix) + W _SDO (n) * G _{pro SDO} (ix) + W _GK (n) * G _{pro GK} (ix),
V _Pro (ju, n) = W _diff (n) * V _{ave DIF} (ju) + W _LMS (n) * V _{direct LMS} (ju) + W _CC (n) * V _{ave HA} (ju)
where n is the number of the current frame. 14. The method of claim. 13, characterized in that the current X coordinate _OBIN, Y _OBIN generalized binary television image of the object is defined as the weighted sum of the coordinates of the center of gravity _bct X, Y coordinates and _{bct OMED} X, Y _OMED median area television generalized binary object image
X _OBIN (n) = W _CT (n) * X _OCT + [1-W _CT (n)] * X _OMED ,
Y _BOTH (n) = W _CT (n) * Y _BTC + [1-W _CT (n)] * Y _OMED ,
moreover, the weight coefficient W _CT (n) of estimating the coordinates of the center of gravity of the generalized binary television image of the object is increased with decreasing average deviation of the coordinates of the television image of the object from their predicted values. 15. The method according to p. 1, characterized in that the current horizontal VG _{OB_BIN} and vertical VV _{OB_BIN} components of the estimated velocity of the generalized binary television image of the object in the inertial coordinate system is determined in accordance with the expressions
VG _{OB_BIN} = (dX + ΔX _OA + ΔX _{OB_OA_BIN} ) / T,
VV _{OB_BIN} = (dY + ΔY _OA + ΔY _{OB_OA_BIN} ) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OB_OA_BIN} , ΔY _{OB_OA_BIN} - change the coordinates of the generalized binary television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively. 16. The method according to p. 1, characterized in that the signals of the dynamic reference television image of the object are formed by reading in each frame the signals from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET Whose} Y _{ET ET} is determined by the difference
X _{C ET} = X _{AB PZ} -X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA PZ} ,
where X _{ABOUT PZ} , Y _{ABOUT PZ} - the coordinates of the television image of the object in the field of view of the video camera of the surveillance system;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system. 17. The method according to claim 1, characterized in that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , Y _{C ET} which is determined by the difference
X _{C ET} = X _{ABOUT PZ} -X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA P3} ,
wherein X _{ON PP,} Y _{ON PP} - the coordinates of the television picture of the object in the surveillance video camera system view;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system;
when the conditions for changing the static reference television image of the object formed by comparing the parameters of the signals of the measures of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static and dynamic reference television images of the object, as well as comparing the parameters of the trajectories of the television image of the object obtained when using the signals static and dynamic reference television images of the object. 18. The method according to p. 1, characterized in that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , Y _{C ET} which is determined by the difference
X _{C ET} = X _{ABOUT PZ} -X _{OA P3} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA PZ} ,
wherein X _{ON PP,} Y _{ON PP} - the coordinates of the television picture of the object in the surveillance video camera system view;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system;
when the conditions for changing the static reference television image of the object, formed on the basis of the analysis of the signal parameters of the measure of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static reference television image of the object, as well as the path parameters of the television image of the object obtained on the basis of signal analysis measures of dissimilarity of television images. 19. The method according to p. 1, characterized in that the current horizontal VG _{OB_NXX} and vertical VV _{OB_NXX} components of the estimation of the speed of movement of the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, is determined in accordance with the expressions
_{OB_NSKH} VG = (dX + ΔX + ΔX _{OA OB_OA NSH)} / T,
_{OB_NSKH} VV = (dY + ΔY _{OA OB_OA NSH} + ΔY) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OB_OA_NXX} , ΔY _{OB_OA_NXX} - change the coordinates of the television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, obtained on the basis of the formation of signals of the measure of dissimilarity of television images. 20. The method according to p. 1, characterized in that the confidence coefficient W _BIN (n) of the current speed of the generalized binary television image of the object is obtained by determining the current density

generalized binary television image, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image, then averaging the limited density of the generalized binary television image with a first order recursive filter, normalizing the average density

generalized binary television image of an object

where S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object;
S _OO (n) is the current area of the region inside the boundaries of the generalized binary television image of the object;
n is the number of the current frame;

- average similarity coefficient obtained by determining the current similarity coefficient

restrictions on its maximum and minimum values and averaging by a first order recursive filter;
σ _{f min} (n) is the minimum rms value of the background television image signals in the background analysis windows;
E _{HCX min} (n) is the minimum value of the signals of the measure of dissimilarity of the television image in the current analysis window and the static reference television image of the object in the two-dimensional region of the search for the displacements of the television image of the object. 21. The method according to p. 18, characterized in that the coefficient W _CX (n) of the current speed of movement of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, is obtained by determining the current coefficient of similarity

limiting its maximum and minimum values and averaging with a first-order recursive filter, normalizing the obtained average similarity coefficient

where n is the number of the current frame;
σ _min (n) is the minimum value of the rms value of the background television image signals in the background analysis windows;
E _HCXmin (n) is the minimum value of the signals of the measure of dissimilarity of the television images of the current analysis window and the static reference television image of the object in the two-dimensional region of the search for the displacements of the television image of the object;

- the average density of the generalized binary television image of the object obtained by determining the current density

generalized binary television image of the object, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image of the object and then averaging, with a first order recursive filter, the limited current density V _BIN (n) of the generalized binary television image of the object;
S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object,
S _OO (n) is the current area of the region inside the boundaries of the generalized binary television image of the object. 22. The method according to p. 1, characterized in that a comprehensive assessment of the horizontal VG _OB (n) and vertical VV _OB (n) components of the current speed of the television image of the object in the inertial coordinate system is determined by limiting the estimates of the speed of the binary television image of the object and speed movement of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, the minimum and maximum values generated taking into account the previous The values of integrated assessment of the current speed of movement of television images of the object and forming a weighted sum of the count rate of movement limited television binary object image and the speed of movement of television images of the object obtained based on the signal generating dissimilarity measure television pictures
VG _OB (n) = W _BIN (n) * VG _{OB_BIN} (n) + W _CX (n) * VG _{ABOUT NCH} (n),
VV _OB (n) = W _BIN (n) * VV _{OB_BIN} (n) + W _CX (n) * VV _{ABOUT NCH} (n),
where W _BIN (n), W _CX (n) are the reliability coefficients of the current velocity of the generalized binary television image of the object and the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, respectively;
VG _{OB_BIN} (n) and VV _{OB_BIN} (n) - limited horizontal and vertical components of the current speed of movement of the generalized binary television image of the object;
VG _{ABOUT NSX} (n) and VV _{ABOUT NSX} (n) - limited horizontal and vertical components of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images;
n is the number of the current frame. 23. The method according to p. 1, characterized in that the analysis of the current S _OBIN (n) and averaged

the area of the generalized binary television image of the object is produced by checking the fulfillment of the conditions:

- to go to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates of the television image of the object or condition,

- to proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window, where ksl is a constant coefficient, ksl <l; ks2 (nn _ek ) - coefficient decreasing with increasing frame number n, starting from frame number n _ek , the transition to the formation of control signals for moving the axis of the camera’s field of view along extrapolated coordinates, ks2 (nn _ek ) ≤ksl. 24. A signal processing device for determining the coordinates of objects observed in a sequence of television images, comprising a block (4) for generating signals of a measure of dissimilarity of television images 2N of terrain landmarks and for determining shift and rotation parameters of signals of the current field of the television image during the time between receiving signals of the current and previous fields television images and a processor (24) for the computational processing of local data in the separation of time on a common bi-directional bus, I distinguish it is that a block (1) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for the operation of the device, a block (2) for receiving and storing signals for uncontrolled movement and roll of the field of view of the camera of the surveillance system, block (3) were introduced calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the previous and current fields of the television image, block (5) separation obtained based on the assessment measures dissimilarities of television images of 2N landmarks of the shift parameters of the signals of the current field of the television image between the reception of signals of the current and previous fields of television images by the components of the controlled movement of the axis of the field of view of the video camera of the monitoring system and the uncontrolled movement of the axis of the field of view of the video camera of the monitoring system, switch block (6), shaper (7) of television image signals of the current frame from the television image signals of the previous frame with the volume of the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals of the previous and current fields of the television image and the signals of the current field of the television image with the elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the monitoring system, shaper ( 8) television image signals in the current analysis window using position and size signals ditch of the current window for analyzing the television image, the shaper (9) of the windows for analyzing the background around the window for analyzing and determining the projections of the signals of the binary television image of the histogram classifier in the windows for analyzing the background, the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of a generalized binary television image of an object, a block (11) for determining the coordinates of a television image of an object the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, the unit (12) for determining the current speed of movement of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object obtained on the basis of the formation of signals of the measure of dissimilarity of television images, switch (13) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, the analyzer (14) of the current speed place a generalized binary television image of an object in an inertial coordinate system, an analyzer (15) of the current speed of a television image of an object in an inertial coordinate system, obtained based on the formation of signals of a measure of dissimilarity of television images, a block (16) for determining the area and coordinates of the boundaries of binary television images of a histogram classifier in the background analysis windows, block (17) for determining the current speed of movement of the generalized binary television image about of an object in an inertial coordinate system, a unit (18) for generating a comprehensive estimate of the current speed of moving the television image of an object in an inertial coordinate system, a unit (19) for evaluating the coordinates of a television image of an object in the field of view of a video camera of an observation system, an analyzer (20) for using extrapolated television coordinates images of the object based on the current and average area of the generalized binary television image of the object, the current and average speed the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the unit (21) for averaging a complex estimate of the current speed of the television image of the object and storing the values of the filtered speed, the extrapolator (22) of the coordinates and speed of the television image of the object in the next frame based on the analysis of the values of the averaged speed of the television image of the object, the shaper (23) signal in controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and the speed of the television image of the object, the first and second inputs of the device are connected with the first and second inputs of the block (1) receiving and storing signals of the current field of the television image from the video camera with monitoring and generating systems for synchronizing the operation of the device, respectively, the third input of the device is connected to the input of the block (2) for receiving and storing signals of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system, the fourth and fifth inputs of the device are connected to the first and second inputs of the switch (13) codes initial or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, respectively, the sixth input of the device is connected to the second input of the calculation unit (3) of the movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals from the previous and current fields of the television image, the seventh input of the device is connected to the fifth input of the block (6) of switches, the first output of the block (1) receiving and storing signals of the current field of the television image from the video camera of the system observation and generation of signals for synchronizing the operation of the device is connected to the first input of the block (4) for generating signals of a measure of dissimilarity of television images 2N landmarks and determined the parameters of the shift and rotation of the signals of the current field of the television image between the reception of signals of the current and previous fields of television images and with the first input of the imager (7) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the surveillance system during the time between the reception of signals of the previous and current fields of the television image and signals of the current field of the television image with oelementnym coordinate transformation signals of the current field of television images, providing
compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, the second output of the unit (1) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device is connected to the third input of the block (3) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between receiving signals from the previous and current fields television image the fifth input of the block (8) for generating and scaling television image signals in the current analysis window using the position and size signals of the current television image analysis window, the fourth input of the shaper (7) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals of the previous and current fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current television image fields providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, the second input of the separation unit (5) of the television images obtained on the basis of the measure of dissimilarity of the 2N landmarks of the shift parameters of the signals of the current field of the television image during the time between the reception of signals of the current and previous fields of the television images on the components of the controlled movement of the axis of the field of view of the video camera of the monitoring system and uncontrolled movement of the axis field of view of a video camera of a surveillance system, the fourth input of the block (4) generating signals of the measure of dissimilarity of television images 2N landmarks and determine the parameters of the shift and rotation of the signals of the current field of the television image during the time between receiving signals of the current and previous fields of television images, the second inputs of the shaper (9) of the background analysis windows around the perimeter of the analysis window and determine the projections of the binary television image signals of the histogram classifier in the background analysis windows and the block (16) for determining the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the third input of the block (10) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, the fourth input of the block (12) determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, the fifth input of the switch (13) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, the fourth inputs of the analyzer (14) of the current velocity of the generalized binary television image of the object in the inertial coordinate system and the analyzer (15) of the current velocity of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the fourth input of the block (17) determining the current velocity of the generalized binary television image of the object in an inertial coordinate system, the fifth input of the block (18) forming a comprehensive assessment of the current speed of movement of the television image of the object in an inertial coordinate system, the third input of the block (19) a comprehensive assessment of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the sixth input of the analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the third inputs of the averaging unit (21) for an integrated assessment of the current speed of moving a television image of an object and storing the values of the filtered speed and extrapolator (22) of coordinates and the speed of moving a television image of an object in the next frame based on the analysis of the values of the average speed of a television image of an object and the sixth input of the shaper ( 23) control signals for moving the axis of the field of view of the video camera of the surveillance system and signals for the position and size of the analysis window in blowing television picture frame using the coordinates of the object in the television picture field of view of video camera surveillance system or extrapolated coordinates and velocity of the television image of the object, the output of the block (2) for receiving and storing signals of uncontrolled movement and the roll of the field of view of the video camera of the monitoring system is connected to the third input of the block (4) for generating signals of the measure of dissimilarity of television images 2N of terrain landmarks and determining the parameters of the shift and rotation of the signals of the current field of the television image during the time between reception signals of current and previous fields of television images and with the third input of the block (6) switches, the first output of the block (3) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between receiving signals from the previous and current fields of the television image is connected to the second input of the block (4) for generating signals of the measure of dissimilarity of the television images 2N of the terrain landmarks and determining the parameters of the shift and rotation of the signals the current field of the television image during the time between receiving signals of the current and previous fields of the television image and with the fourth input of the block (6) of the switch her, the first output of the block (4) for generating signals of a measure of dissimilarity of television images of 2N landmarks and determining the parameters of the shift and rotation of the signals of the current field of the television image during the time between receiving signals from the current and previous fields of television images is connected to the first input of the separation unit (5) obtained based on the estimate measures of dissimilarity of television images 2N landmarks of the shift parameters of the signals of the current field of the television image in the time between receiving signals of the current and previous fields of television images on the components of the controlled movement of the axis of the field of view of the video camera of the surveillance system and the uncontrolled movement of the axis of the field of view of the video camera of the surveillance system, the first and second outputs of which are connected to the first and second inputs of the block (6) of the switches, the first output of the block (6) of switches is connected to the second input of the shaper (7) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals from the previous and current fields of the television image and signals of the current fields of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, and the second output is connected to the third input of the shaper (7) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals of the previous and current fields of the television image and signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, with the fifth input of the block (10) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the first input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, with the first input of the block (12) determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, with the second input of the block (17) determining the current speed of the generalized binary television image of the object in an inertial coordinate system, and with the first input of the block (19) of an integrated assessment of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the output of the imager (7) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the signals of the previous and current fields of the television image and the signals of the current field of the television image with elementwise coordinate transformation of the signals of the current field television image providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, connected to the first input of the imager (8) of the television image signals in the current analysis window using the position and size signals of the current television image analysis window, the output of the imager (8) of the television image signals in the current analysis window using the position and size signals of the current television image analysis window is connected to the seventh input of the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image object and with the third input of the block (11) determining the coordinates of the television image of the object relative to the center t a checking window analysis on the basis of a signal generating dissimilarity measure television images, the output of the shaper (9) of the background analysis windows along the perimeter of the window for analyzing and projecting the signals of the binary television image of the histogram classifier in the background analysis windows is connected to the first input of the block (16) for determining the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the first output of which is connected to the third input of the analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the first output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the shaper (9) of the background analysis windows along the perimeter of the analysis window and for determining the projections of the signals of the binary television image of the histogram classifier in the background analysis windows, the second output of the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the switch (13) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, the first input of the analyzer (14), the current speed of movement of the generalized binary television image of the object in the inertial coordinate system and the first input of the shaper (23) of the signals to control the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera systems of observation or extrapolated coordinates and speed of movement of the television image of the object, the third output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the block (17) for determining the current speed of the movement of the generalized binary television image of the object in an inertial coordinate system, the fourth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the second input of the analyzer (14) of the current speed of the movement of the generalized binary television image of the object in the inertial coordinate system and to the second input analyzer (20) conditions for the use of extrapolated coordinates of the television image of the object based on those uschey and averaged area generalized binary television image of the object, current and average speed of movement of the television image, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the fifth output of the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the analyzer (20) for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image images of the object the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the sixth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the analyzer (15) of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the seventh output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the fourth input of the block (1) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating synchronization signals device operation the first output of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the second input of the block (12) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the third input of the block (18) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the second output of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the first input of the analyzer (15) of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the second input of the block (18) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the first output of the switch (13) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the first input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the seventh input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the third input of the shaper (8) of the television image signals in the current analysis window using the position and size signals of the current analysis window of the television image, the second output of the switch (13) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the second input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the sixth input of the unit (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the second input of the imager (8) of the signals of the television image in the current analysis window using the position and size signals of the current analysis window of the television image, the first output of the analyzer (14) of the current velocity of the generalized binary television image of the object in the inertial coordinate system is connected to the first input of the block (18) for generating a complex estimate of the current speed of the television image of the object in the inertial coordinate system, and the second output is connected to the second input of the analyzer (15) of the current speed of the television image of the object in an inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the second output of which is connected to the third input of the analyzer (14) of the current velocity of the generalized binary television image of the object in an inertial coordinate system, the first output of the block (17) for determining the current moving speed of the generalized binary television image of the object in the inertial coordinate system is connected to the fourth input of the block (18) for generating a complex estimate of the current speed of moving the television image of the object in the inertial coordinate system, the first output of the unit (18) for generating a comprehensive estimate of the current speed of the television image of the object in the inertial coordinate system is connected to the fourth input of the analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the fifth input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, the first input of the unit (21) for averaging the complex estimate of the current speed of the television image of the object and storing the values of the filtered speed and the second input of the unit (19) for the complex estimate of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the first output of the unit (19) for the complex assessment of the coordinates of the television image of the object in the field of view of the video camera of the monitoring system is connected to the fourth input of the switch (13) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, the first input of the extrapolator (22) coordinates and the speed of the television image of the object in the next frame based on the analysis of the values of the average speed of the television image of the object and the third input of the shaper (23) of the signals to control the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of a video camera of a surveillance system or ex rapolirovannyh coordinates and velocity of the television image of the object, the second output of which is connected to the third inputs of the block (12) for determining the current speed of the television image of the object in an inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of the formation of the signals of the measure of dissimilarity of the television images and the block (17) for determining the current speed of movement of the generalized binary television image of the object in an inertial coordinate system, the sixth input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object and the fourth inputs of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of television dissimilarity imaging and imaging unit (8) of the television image signals in the current analysis window using m signals of the current position and dimensions of the television image analysis window, the first output of the analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows is connected to the second input of the generator (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field view of the video camera of the surveillance system or extrapolated coordinates and the speed of movement of the television image of the object, the second input of the unit (21) for averaging a complex estimate of the current speed of moving the television image of the object and storing the values of the filtered speed and with the fourth input of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the second output of the analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows is connected to the second input of the unit (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images, the fourth output of the analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows is connected to the third input of the unit (1) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device, the first output of the unit (21) for averaging a complex estimate of the current speed of moving the television image of the object and storing the values of the filtered speed is connected to the fifth input of the analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows and the second input of the extrapolator (22) coordinates and the speed of the television image of the object in the next frame based on the analysis of the values of the average speed of the television image of the object, the second output of the unit (21) for averaging a complex estimate of the current speed of moving the television image of the object and storing the values of the filtered speed of the object is connected to the fifth input of the generator (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system or extrapolated coordinates and speed ty moving the television image of the object, the first output of the extrapolator (22) coordinates and the speed of the television image of the object in the next frame, based on the analysis of the values of the average speed of the television image of the object, is connected to the fourth input of the driver (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera system is observed or extrapolated coordinates and speed of the television image of the object, the first output of the driver (23) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and the speed of movement of the television image of the object is the output devices and is connected to the first input of the block (3) for calculating the controlled movement of the axis of the field of view of the video camera systems s monitoring the time between the reception of signals from the previous and current fields of the television image, the output of the processor (24) for computing local data in a time-sharing shared bi-directional bus connected to the second outputs of the unit (3) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals from the previous and current fields of the television image, unit (4) for generating signals of a measure of dissimilarity of television images 2N landmarks and determining the parameters of the shift and rotation of the signals of the current field of the television image during the time between receiving signals of the current and previous fields of television images, unit (12) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, block (16) determining the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, block (17) determining the current speed of the generalized binary television image of the object in an inertial coordinate system, block (18) forming a comprehensive assessment of the current speed of the television image of the object in an inertial coordinate system, block (19) a comprehensive assessment of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, extrapolator (22) coordinates and speed of the television image of the object in the next frame based on the analysis of the values of the average speed of the television image of the object, the third outputs of the analyzer (14) of the current velocity of the generalized binary television image of the object in an inertial coordinate system, analyzer (15) the current speed of the television image of the object in an inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, analyzer (20) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, block (21) of averaging a comprehensive estimate of the current speed of the television image of the object and storing the values of the filtered speed and the driver (23) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of a video camera of a surveillance system or extrapolated coordinates and speed of movement of a television image facility the eighth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object.  25. The device according to p. 24, characterized in that the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object contains a switch (31) of video data, the first buffer random access memory (30 ), a second buffer random access memory (41) and a third buffer random access memory (43), a shaper (32) of signals of a binary television image r the histogram classifier in the current analysis window, the shaper (36) of the signals of the primary binary television image of the detector of changes in the background, the shaper (38) of signals of the secondary binary television image of the detector of changes in the background, the shaper (45) of signals of the primary binary television image of the selector of moving objects, the shaper (47) signals of the secondary binary television image of the selector of moving objects, the first node (42) scaling and shifting the signals of the television from images, node (44) for generating signals of a differential television image of a selector of moving objects, shaper (33) of horizontal and vertical projections of signals of a binary television image of a histogram classifier, shaper (39) of horizontal and vertical projections of signals of a secondary binary television image of a detector of background changes, shaper (48 ) horizontal and vertical projections of the signals of the secondary binary television image of the selector of moving objects, b lock (35) for determining the object / background ratio and the minimum rms value of the background television image signals in the background analysis windows, a spatial filter (37) for the low frequencies of the primary binary television image signals of the background change detector, a spatial filter (46) for the low frequencies of the primary binary television image signals moving object selector, analyzer (34) of signals of binary television images of a histogram classifier, analyzer (40) of signals of secondary binary television images of the background change detector, an analyzer (49) of signals of secondary binary television images of the selector of moving objects and a generator (50) of generalized horizontal and vertical projections of signals of the generalized binary television image of the object, a computer (51) of the coordinates and dimensions of the object, the current and average area of the generalized a binary television image of an object, and an analyzer (52) of conditions for the breakdown of automatic determination of the coordinates of a television image of an object, m the first and second inputs of the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object are connected to the first and second inputs of the generator of the binary television image of the histogram classifier in the current analysis window and unit (35) for determining the object / background ratio and the minimum rms value of the signals of the television background image in the background analysis buttons and the fourth and third inputs of the primary binary television image signal generator (36) of the background change detector, respectively, the first input of the unit (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is also connected with the first input of the first node (42) scaling and shifting the signals of the television image, the third input of the block (10) is determined The current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object are connected to the third input of the video data switch (31), the seventh input of the histogram binary television image signal (32) signal in the current analysis window, and the second input of the generator (33) horizontal and vertical projections of signals of a binary television image of a histogram class ikator, the second input of the analyzer (34) of signals of binary television images of the histogram classifier, the fifth input of the block (35) for determining the ratio of object / background and the minimum rms value of the signals of the television image in the background analysis windows, the sixth input of the imager (36) of the signals of the primary binary television image detector of changes in the background, the second input of the analyzer (40) of the signals of the secondary binary television images of the detector for changes in the background, the fourth input of the first node (42 ) scaling and shifting the signals of the television image, the second input of the analyzer (49) of the signals of the secondary binary television images of the selector of moving objects, the seventh input of the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object and the second input of the calculator (51) of coordinates and sizes object, the current and average area of the generalized binary television image of the object, the fourth input of the block (10) determine the current coordinates of of a generalized binary television image of an object using generalized horizontal and vertical projections of a generalized binary television image of an object is connected to the fourth input of the histogram classifier binary signal image generator (32) in the current analysis window, the fifth input of the block (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of a generalized binary television of the object image is connected to the first input of the primary binary television image signal generator (36) of the background change detector and the second input of the first television image scaling and shifting node (42), the sixth input of the unit (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of a generalized binary television image of the object is connected to the second input of the signal shaper (36) ne the primary binary television image of the background change detector, the fifth input of the histogram classifier binary signal image generator (32) in the current analysis window and the third input of the block (35) for determining the object / background ratio and the minimum rms value of the background image television signals in the background analysis windows, seventh block input (10) for determining the current coordinates of a generalized binary television image of an object using generalized horizontal and vertical The projection of the generalized binary television image of the object is connected to the inputs of the first buffer random access memory (30) and the second buffer random access memory (41), the first input of the differential television image generating unit (44) of the signals of the moving objects selector, the fifth input of the signal shaper (36) the primary binary television image of the background change detector and the first input of the first video data switch (31), the output of the first operational buffer the blinking device (30) is connected to the second input of the video data switch (31), the first and second outputs of which are connected to the third and sixth inputs of the signal generator (32) of the binary television image of the histogram classifier in the current analysis window, respectively, and with the fourth input of the block (35) determining the object / background ratio and the minimum rms value of the background television image signals in the background analysis windows, the first output of which is the fifth output of the current coordinate determining unit (10) the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the first output of the binary television image signal generator (32) of the histogram classifier in the current analysis window is connected to the first output of the unit (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of a generalized binary TV the image of the object and with the first input of the shaper (33) of the horizontal and vertical projections of the signals of the binary television image of the histogram classifier, the output of which is connected to the first input of the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object and the first input of the analyzer (34 ) signals of binary television images of a histogram classifier, the first output of which is connected to the second input of the shaper (50) of horizontal and vertical projections of the signals of the generalized binary television image of the object, the output of the shaper (36) of the signals of the primary binary television image of the background change detector is connected to the input of the spatial filter (37) of the low frequencies of the signals of the primary binary television image of the background change detector, the output of which is connected to the input of the shaper (38) signals of the secondary binary television image of the background change detector, the output of which is connected to the input a shaper (39) of horizontal and vertical projections of the signals of the secondary binary television image of the background change detector, the output of which is connected to the third input of the shaper (50) of the generalized horizontal and vertical projections of signals of the generalized binary television image of the object and with the first input of the analyzer (40) of the signals of the secondary binary television images of the detector of changes in the background, the first output of which is connected to the fourth input of the generator (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the output of the second buffer random access memory (41) is connected to the third input of the first node (42) of scaling and shifting the signals of the television image, the output of which is connected to the input of the third buffer random access memory (43), the output of which connected to the second input of the node (44) for generating signals of the differential television image of the moving objects selector, the output of which is connected to the input of the shaper (45) signals of the primary binary television image of the moving object selector, the output of which is connected to the input of the spatial filter (46) low frequencies of the signals of the primary binary television image of the moving object selector, the output of which is connected to the input of the generator (47) of signals of the secondary binary television image of the moving object selector the output of which is connected to the input of the shaper (48) of horizontal and vertical projections of the signals of the secondary binary television images of the moving object selector, the output of which is connected to the first input of the analyzer (49) of the signals of the secondary binary television images of the selector of moving objects and the sixth input of the generator (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the first output of the analyzer (49) of the signals of the secondary binary television images of the moving object selector is connected to the fifth input of the shaper (50) of generalized horizontal and vertical of the signal signals of the generalized binary television image of the object, the first output of which is connected to the first input of the computer (51) of the coordinates and dimensions of the object, the current and average area of the generalized binary television image of the object and with the fourth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of a generalized binary television image of an object, the first and second outputs of the computer (51) coordinates and dimensions of the object, the current and average area of the generalized binary television image of the object are the second and third outputs of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, respectively, the third output of the calculator (51 ) coordinates and dimensions of the object, the current and average area of the generalized binary television image This is connected to the input of the analyzer (52) of the conditions for the breakdown of automatic determination of the coordinates of the television image of the object, the first output of which is the seventh output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the third output unit (35) for determining the object / background ratio and the minimum rms value of the television image signals The background in the background analysis windows is connected to the sixth output of the block (10) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth output of which is connected to the second outputs of the histogram binary television image signal (32) classifier in the current analysis window, block (35) for determining the object / background ratio and the minimum rms value background television image signals in the background analysis windows, the analyzer (34) of the binary television image signals of the histogram classifier, the analyzer (40) of the secondary binary television image signals of the background change detector, the analyzer (49) of the secondary binary television image signals of the moving object selector, analyzer (52) failure conditions for automatic determination of the coordinates of the television image of the object and the shaper (50) of the generalized horizontal and vertical projections of signals about television communication binary object image and the fourth output of the calculator (51) coordinates and size of the object, and the current average area of a generalized binary television image object.  26. The device according to p. 25, characterized in that the generator (32) of the signals of the binary television image of the histogram classifier in the current analysis window contains a calculator (67) normalized histograms of the distribution of the brightness of the signals of the television images of the object and background and the node (68) of the formation of the binary television image histogram classifier, and, the first, second, third, fourth and fifth inputs of the shaper (32) of signals of the binary television image of the histogram classifier in the current not analyzed are connected to the first, second, third, fourth and fifth inputs of the calculator (67) of the normalized histograms of the distribution of the brightness of the signals of the television images of the object and background, respectively, the sixth input is connected to the second input of the node (68) of the formation of the binary television image of the histogram classifier, the seventh input of the shaper (32) signals of the binary television image of the histogram classifier in the current analysis window is connected to the sixth input of the calculator (67) normalized histograms of the distribution the luminance of the signal brightness of the television images of the object and background and the third input of the node (68) forming the binary television image of the histogram classifier, the first output of the calculator (67) normalized histograms of the distribution of the brightness of the signals of the television image of the object and background is connected to the first input of the node (68) of the formation of the binary television image histogram classifier, the output of which is the first output of the generator (32) of signals of the binary television image of the histogram assifikatora in the current window of analysis, the second output of the calculator (67) of the normalized histogram distribution of the brightness signals of television pictures and the background object is connected to the output of the second (32) binary signals television image histogram classifier in the current analysis window.  27. The device according to p. 25, characterized in that the imaging unit (36) of the signals of the primary binary television image of the detector for changing the background contains the imaging unit (70) of the signals of the television image of the reference background, the first (69) and the second (72) buffer random access memory, node (71) scaling and shifting the signals of the television image of the reference background, the node (73) for generating the signals of the differential television image of the detector for changing the background and the node (74) for the primary binarization of the detector for changing the background, and, the first and second inputs of the primary binary television image signal generator (36) of the background change detector are connected to the first and second inputs of the television image signal generator (70) of the background image and the node (71) of scaling and shifting the signals of the television image of the background background, respectively, the third and fourth inputs connected to the third and fourth inputs of the shaper (70) of the television image signals of the reference background, respectively, and the fifth input is connected to the first input of the node (73) of images signals of the differential television image of the background change detector and the input of the first buffer random access memory (69), the output of which is connected to the fifth input of the signal generator of the television image signal (70) and the sixth input of the signal generator (36) of the primary binary image signal of the background change detector connected to the third input of the node (73) for generating the difference television image of the background change detector and the seventh input of the signal shaper (70) t a levision image of a reference background, the output of which is connected to the third input of the scaling and shifting signals node (71) of the television image of the reference background, the output of which is connected to the input of the second buffer random access memory (72), the output of which is connected to the sixth input of the signal generator (70) of the television images of the reference background and the second input of the node (73) for generating the signals of the differential television image of the background change detector, the output of which is connected to the input of the node (74) of the primary the background of the detector for changing the background, the output of which is connected to the output of the former (36) of the signals of the primary binary television image of the detector for changing the background. 28. The device according to p. 24, characterized in that the unit (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images contains a switch (53) of signals of a scaled image, a non-linear high-pass filter (54) , an analyzer (55) of conditions for updating a static reference television image of an object, an analyzer (61) of the type of sequences of minimum values of signals of a measure of dissimilarity of television images to Only rows and along the columns of the two-dimensional region for searching for displacements of the television image of the object, the first (56) and second (59) buffer random access memory devices, the shaper (57) of the signals of the static and dynamic reference television images of the object, the shaper (60) of signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for displacements of the television image of the object, the node (58) for scaling the signals of the television image, the selector (62) of the sequence of minimum values the signals of the measure of dissimilarity of television images, the approximator (63) of the sequence of the minimum values of the signals of the measure of dissimilarity of television images by a polynomial of the fourth degree, the coordinator (64) of the television image of the object in the analysis window according to the minimum position of the approximating polynomial, the coordinator (65) of the television image of the object in the analysis window by offset boundaries of the region of rapid growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window, coordinator (66) of television of the image of the object in the analysis window according to the coordinates of the center of the analysis window, and the first input of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the first inputs of the analyzer (55) for updating the static reference of a television image of an object and a shaper (60) of signals of a measure of dissimilarity between signals of television images in the two-dimensional domain of the second image of the object, the second input of the block (11) of determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the first input of the switch (53) of the signals of the scaled image, the third input of the block (11) of determining the coordinates of the television image object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images is connected to the second input of the switch (53) signals of the scaled image and with the first input of the nonlinear high-pass filter (54), the output of which is connected to the third input of the switch (53) of the signals of the scaled image, the fourth input of the block (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on signal generation measures of dissimilarity of television images is connected to the third input of the shaper (60)
signals of a measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for the displacements of the television image of the object and with the second inputs of the shaper (57) of the signals of the static and dynamic reference television images of the object and the analyzer (55) of the conditions for updating the static reference television image of the object, the fifth input of the block (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television of these images is connected to the fourth input of the imaging unit of signals (60) of the dissimilarity between the signals of the television images in the two-dimensional region for searching for displacements of the television image of the object, the sixth input of the unit (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of the dissimilarity of the television images connected to the third input of the shaper (57) of the signals of the static and dynamic reference television images of the object, the seventh input One of the unit (11) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the fourth input of the shaper (57) of the signals of the static and dynamic reference television images of the object and to the fifth input of the shaper (60) of the signals of the dissimilarity between the signals of the television images in the two-dimensional area of the search for the displacements of the television image of the object, the eighth input of the block (11) determining the coordinates of a live image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the second input of the nonlinear high-pass filter (54), to the fifth input of the analyzer (55) of the update conditions of the static reference television image of the object, the fourth input of the selector (62) of the sequence the minimum values of the signals of the measure of dissimilarity of television images and the seventh input of the shaper (60) of signals of the measure of dissimilarity between the signals of the television x images in the two-dimensional region for searching for the displacements of the television image of the object, the output of the switch (53) of the signals of the scaled image is connected to the second input of the shaper (60) of signals of dissimilarity between the signals of the television images in the two-dimensional region for searching for the displacements of the television image of the object and the input of the first buffer random access memory ( 56), the output of which is connected to the first input of the shaper (57) of the signals of the static and dynamic reference television images kta, the first output of the analyzer (55) for updating the static reference television image of the object is connected to the fifth input of the shaper (57) of signals of the static and dynamic reference television image of the object, the output of which is connected to the input of the node (58) for scaling the signals of the television image, the output of which is connected to the input of the second buffer random access memory (59), the output of which is connected to the sixth input of the driver (60) of the signals of the measure of dissimilarity between the signals of the television x images in the two-dimensional region for searching for the displacements of the television image of the object, the output of which is connected to the input of the analyzer (61) of the type of sequences of the minimum values of signals of the measure of dissimilarity of the television images along the rows and along the columns of the two-dimensional region for searching for the displacements of the television image of the object, the first output of which is connected to the second output (11) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of measure dissimilarity TV images, the third input of the analyzer (55) for updating the static reference television image of the object and the input of the selector (62) of the sequence of minimum values of the signals of the measure of dissimilarity of television images, the first output of which is connected to the input of the approximator (63) of the sequence of the minimum values of signals of the measure of dissimilarity of television images a polynomial of the fourth degree, the first output of which is connected to the input of the coordinator (64) of the television image of the object in the window a according to the minimum position of the approximating polynomial, the second and third outputs of the selector (62) of the sequence of minimum values of the signals of the measure of the dissimilarity of the television images are connected to the inputs of the coordinator (65) of the television image of the object in the analysis window on the shift of the border of the region of rapid growth of the values of the signals of the measure of dissimilarity of the television images relative to the center the analysis window and the coordinator (66) of the television image of the object in the analysis window according to the coordinates of the center of the analysis window, the first outputs of the coordinates torus (64) of the television image of the object in the analysis window by the position of the minimum of the approximating polynomial, the coordinator (65) of the television image of the object in the analysis window by the shift of the boundary of the region of rapid growth of signal values of the measure of dissimilarity of the television images relative to the center of the analysis window and the coordinator (66) of the television image of the object in the analysis window, the coordinates of the center of the analysis window are connected to the first output of the block (11) for determining the coordinates of the television image of the object relative to the center of the current of the analysis window based on the formation of signals of the measure of dissimilarity of television images and with the fourth input of the analyzer (55) of the conditions for updating the static reference television image of the object, the third output of the block (11) of determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television image is connected to the second outputs of the analyzer (61) of the type of sequences of the minimum values of the signals of the measure of dissimilarity of television and images along the rows and along the columns of the two-dimensional region for searching for the displacements of the television image of the object, the analyzer (55) of the conditions for updating the static reference television image of the object, the approximator (63) of the sequence of the minimum values of the signals of the measure of dissimilarity of the television images by a fourth degree polynomial, the coordinator (64) of the television image of the object in analysis window by the position of the minimum of the approximating polynomial, coordinator (65) of the television image of the object in the analysis window by offset the boundary of the region of rapid growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window and the coordinator (66) of the television image of the object in the analysis window according to the coordinates of the center of the analysis window.  29. A method of processing signals to determine the coordinates of objects observed in a sequence of television images, which consists in receiving and storing signals of the current field of the television image, characterized in that the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals of the previous and current fields of the television image, due to the action on the video camera of the monitoring system of the signals for controlling the movement of its field of view, determine before receiving the signal s of the current field of the television image, determine the speed of the controlled movement of the axis of the field of view of the video camera of the monitoring system from the data of the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of signals of the previous and current fields of the television image, simultaneously with the reception of signals of the current field of the television image, signals are received and stored uncontrolled movement and roll of the field of view of the video camera of the surveillance system and use them to generate signals of the television image of the current frame, the television image signals of the current frame are generated from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view during the time between the reception of the signals of the previous and current fields of the television image and the signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image providing compensation for uncontrolled movements and the roll of the field of view of the CCTV camera observation methods generate and scale the television image signals in the current analysis window using the position and size signals of the current television image analysis window obtained as a result of processing the television image in the previous frame or using extrapolated coordinates and speed of the object, with the initial conditions of the position signals and sizes of the current windows for analyzing a television image formed at the moment of starting tracking an object based on an external START / STOP signal, the signals of the scaled television image in the current analysis window are omitted, the differential television image signals of the moving object selector are generated by subtracting the signals of the scaled television image from the signals of the scaled television image in the current analysis window, previously stored and reduced to the current scale of the television image in the analysis window and shifted by the amount of movement axis of the field of view of the video camera of the surveillance system, form the signals of the primary binary of a television image of a moving object selector from signals of a differential television image of a moving object selector, signals of a secondary binary television image of a moving object selector are generated from signals of a primary binary television image of a moving object selector that have passed low-pass filtering, simultaneously with storing signals of a scaled television image in the current analysis window, the formation of signals of differential television iso the selector moving objects and generating the signals of the primary and secondary binary television images of the moving objects selector from the signals of the scaled television image in the current analysis window and taking into account the signals of the controlled movement of the field of view axis of the video camera of the surveillance system form the signals of the primary binary television image of the background change detector and binary television signals histogram classifier images in the current analysis window, from signals the primary binary television image of the detector of background changes that have passed low-pass filtering, the signals of the secondary binary television image of the detector of background changes are generated, horizontal and vertical projections of the signals of the secondary binary television images of the moving object selector and the detector of background changes, and horizontal and vertical projections of the signals of the binary television image of the histogram classifier , determine the reliability coefficients of the signals secondary binary television images of the background change detector and the moving object selector, as well as signals of the binary television image of the histogram classifier, form generalized horizontal and vertical projections of the signals of the generalized binary television image of the object from the horizontal and vertical projections of the signals of the secondary binary television images of the selector of moving objects and the background change detector , as well as from horizontal and vertical projections of bi of a histogram television image of a histogram classifier based on their joint processing using reliability coefficients of signals of binary television images of a histogram classifier, signals of secondary binary television images of a background change detector and a selector of moving objects, determine horizontal and vertical borders, as well as the dimensions of a television image of an object by cutoff levels on the left and on the right, above and below the set percentage of the area of generalized horizontally oh and vertical projections of the signals of the generalized binary television image of the object, determine the current and average area of the generalized binary television image of the object located inside the formed boundaries of the television image of the object, determine the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object determine the current speed of movement of a binary television image of an object in an inertial coordinate system, a confidence coefficient is determined for the current velocity of a generalized binary television image of an object in an inertial coordinate system, simultaneously with the formation of signals of primary and secondary binary television images of a moving object selector and a background change detector, a binary television image of a histogram classifier, as well as generalized horizontal and vertical projections of signals about the common binary television image of the object is determined by the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of a measure of dissimilarity of the television images as a result of nonlinear high-frequency filtering of the signals of the scaled television image in the current analysis window, provided that the average area of the generalized binary television image of the object exceeds the threshold value , remembering received signals, signals of the static reference television image of the object or signals of the static and dynamic reference television images of the object, bringing the signals of the static reference television image of the object or signals of the static and dynamic reference television images of the object to the current scale, the formation and storage of signals of the measure of dissimilarity between the signals of the television images after non-linear high-frequency filtering signals scaled TV of the image in the current analysis window and the signals of the static or static and dynamic reference television images of the object in the two-dimensional region of searching for the displacements of the television image of the object, determining the minimum values of the signals of the measure of dissimilarity of the television images along the rows and along the columns of the two-dimensional region of searching for the displacements of the television images of the object, forming sequences of minimal signal values of the measure of dissimilarity of television images along rows and along columns the two-dimensional domain for searching for displacements of the television image of the object, determining the corresponding coordinate of the television image of the object in the analysis window, depending on the type of sequence of the minimum values of the signals of the measure of dissimilarity of television images for this coordinate, namely, by analytical approximation of the sequence of the minimum values of signals of the measure of dissimilarity of television images by a fourth degree polynomial and determining the coordinates of the television image of the object, as the minimum approximating polynomial provided that the sequence of minimum values of the signals of the measure of dissimilarity of television images is assigned to the type of sequences with two boundaries of areas of rapid growth of the values of signals of the measure of dissimilarity of television images near the position of its minimum, or by determining the offset of the boundary of the area of rapid growth of values of signals of the measure of dissimilarity of television images relative to the center of the analysis window and the formation of the coordinates of the object in the analysis window as in a value proportional to the obtained shift of the boundary of the region of rapid growth of the signals of the measure of dissimilarity of television images, provided that the sequence of the minimum values of the signals of the measure of dissimilarity of television images is classified as sequences with a flat neighborhood of the minimum position and the presence of one region of rapid growth of the values of the signals of the measure of dissimilarity of television images the coordinates of the object, equal to the coordinate of the center of the analysis window, provided that n the sequence of the minimum values of the signals of the measure of the dissimilarity of the television images to the type of sequences with a flat neighborhood of the position of the minimum values of the signals of the measure of the dissimilarity of the television images in the entire search area of the displacements of the television image of the object, determine the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object obtained based on the formation of signals of a measure of dissimilarity of television from formations, determine the reliability coefficient of the current speed of moving the television image of the object in the inertial coordinate system, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, form a comprehensive assessment of the current speed of moving the television image of the object in the inertial coordinate system from the data of the estimation of the current speed of moving the television image of the object based on the formation of signals of the measure of dissimilarity of television images and The current speed of the object’s movement over the generalized projections of the signals of the generalized binary television image of the object, taking into account the reliability coefficients of the components of the speed of the television image of the object and a priori restrictions on the speed of maneuvering of the object, average the signals of the integrated estimation of the speed of the television image of the object in the inertial coordinate system and store them, determine the coordinates of the television image of the object in the field of view of the camera system observations by integrating the difference of the complex estimate of the current speed of moving the television image of the object in the inertial coordinate system and the speed of the controlled movement of the axis of the field of view of the video camera of the observation system in the inertial coordinate system with the initial coordinate conditions and the speed of movement of the television image of the object, formed at the time of starting tracking the object based on the external signal START / STOP, simultaneously with the formation of signals of primary and secondary binary television and the image of the moving object selector and the background change detector, the binary television image of the histogram classifier, the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object form the background analysis windows along the perimeter of the analysis window and determine the signal projections of the binary television image of the histogram classifier in the background analysis windows, determine the area of the analysis and coordinates of the boundaries of binary television images of a histogram classifier and in the background analysis windows from the obtained projections of the signals of the binary television image of the histogram classifier in the background analysis windows, control signals for the movement of the axis of the field of view of the video camera of the monitoring system are generated using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system, obtained as a result of processing the television image in the current the analysis window or using the extrapolated coordinates and the speed of movement of the television image of the object depending bridges from the results of the analysis of the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, and the extrapolated speed of the television image of the object is formed based on the analysis of the stored values averaged integrated assessment of the speed of movement of the television image of the object, shape signals of the position and size of the current analysis window of the television image for the next frame using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system obtained as a result of processing the television image in the current analysis window or using the extrapolated coordinates and speed of the television image of the object with the initial conditions, formed at the moment of tracking the object based on an external signal START / STOP. 30. The method according to p. 29, characterized in that the controlled movement dx [n] and dy [n] of the axis of the field of view of the camera horizontally and vertically, respectively, is determined by calculating the convolution of control signals X _yp [i], Y _yp [i] moving the field of view of the video camera of the surveillance system with impulse characteristics hx [i] and hy [i] of its drives

where n is the number of the current half frame;
i is the element number of the impulse response;
K is the length of the impulse response. 31. The method according to p. 29, characterized in that the generation of signals L _FRAME (ix, jy) of the television image of the current frame when interlaced from signals L _P-FRAME (i, p, npk) of the current half-frame of the television image and television image L _{- 1KADR} (ix, jy) of the previous frame is carried out by predicting the signals L _FRAD (ix, jy) of the television image of the current frame by shifting the television image L _-1KADR (ix, jy) of the previous frame by the amount of controlled movement dx, dy of the field of view of the video camera of the surveillance system by horizontally and vertically, respectively, during the time between receiving half-frames of a television image
L _KADR (ix, jy) = L _-1FRAME (ix-dx, jy-dy),
and replacing the points of the television image of the current frame with the points of the television image of the current half frame with compensation for the current uncontrollable displacements r _x (p), r _y (p) and roll φ (p) of the field of view of the video camera of the surveillance system
L _KADR [ix (i, p, npk), jy (i, p, npk)] = L _P-KADR (i, p, npk),
where i is the number of the element in the line of the television image of the current half frame,
i = 1,. . , Nk;
ix is the number of the element in the line of the television image of the current frame,
ix = l,. . . , Nk;
p is the line number in the half frame,

jy is the line number in the frame, jy = 1,. . . , Mk;
Nк is the number of elements of a television image in a row;
Mk - the number of lines in the frame of the television image;
npk - index of the current half-frame:
npk = 1 - in odd half frames,
npk = 0 - in even half frames,

32. The method according to p. 29, characterized in that the signals L _{1BIN DIF} (ix, jy) of the primary binary television image of the background change detector are formed by converting the signals L _{n-1 eff} (ix, jy) of the television image of the reference background obtained in the previous n -1 frame, to the current scale, by generating the signals L _{p DIF} (ix, jy) of the differential television image of the background change detector by subtracting the scaled television image from the signals L _{n OA} (ix, jy) in the current signal analysis window L _{n-1 eff} (ix , jy) television image reference background previous frame shift, which takes into account movement of Vx, Vy analysis window center in the inertial frame for the last frame:
L _{p DIF} (ix, jy) = L _{n OA} (ix, jy) - L _{n-1 eff} (ix + Vx, jy + Vy),
determining the binarization threshold of the _DIF threshold (ix, jy) of the detector of background changes as a value proportional to the local parameter of the scattering of the values of the signals of the differential television image L _pDIF (ix, jy) of the detector of background changes in the vicinity of the point with coordinates ix, jy, assigning values to the primary binary television image L _{1BIN DIF} (ix, jy) of the background change detector

or

moreover, the signals L _{n eff} (ix, jy) of the television image of the reference background are formed by dividing the signals of the scaled television image in the current analysis window into three types of television image signals: the television image signals in the object window are OO _differential , the television image signals in the background window are OF _DIF , the signals of the television image in the window - "New background" - NF, where as the signals of the television image in the window of the object determine the signals of the television image in the rectangle located the center of the current analysis window and including mainly the image elements of the object, as the signals of the television image of the "New background" window, image elements are defined at the outer borders of the current analysis window, on which new elements appear due to the movement of the object and the movement of the axis of the field of view of the video camera of the surveillance system the television image of the background, as the signals of the television image of the background window, all the remaining elements of the television image of the analysis window are determined, storing m in the "New background" window of the signals L _nОA (ix, jy) from the current analysis window of the scaled television image of the current nth frame:
L _{n eff} (ix, jy) = L _nOA (ix, jy), ix, jy ∈ NF,
where L _nOA (ix, jy) are the values of the brightness signals of the element of the scaled television image in the analysis window with the coordinates ix, jy,
averaging in the background window the signals of the scaled television image L _nOA (ix, jy) from the current analysis window with a constant Wof and taking into account the shift of the analysis window in the inertial coordinate system for the last frame:
L _{n eff} (ix, jy) = (1-Wof) * L _{n-1 eff} (ix + Vx, jy + Vy) + Wof * L _nOA (ix, jy), ix, jy ∈ OFP _DIF ,
where Vx, Vy - moving the center of the analysis window for the last frame horizontally and vertically, respectively
in an inertial coordinate system, overwriting in the window of the signal object of the television image the reference background of the previous frame with a shift that takes into account the movement of the center of the analysis window for the last frame:
L _{n eff} (ix, jy) = L _{n-1 eff} (ix + Vx, jy + Vy), ix, jy ∈ OO _DIF . 33. The method according to p. 32, characterized in that the low-pass filtering of the signals of the primary binary television image L _{1BIN DIF} (ix, jy) detector changes in the background is carried out using two-dimensional convolution

where NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
h _HF [di, dj] - impulse response of a low-pass filter. 34. The method according to p. 29, characterized in that the low-pass filtering of the signals of the primary binary television image L _{1 BIN SDO} (ix, jy) of the selector of moving objects is carried out using two-dimensional convolution

where NF, MF are the aperture parameters of the low-pass filter horizontally and vertically, respectively;
h _LF [di, dj] - impulse response of a low-pass filter. 35. The method according to p. 33, characterized in that the signals L _{2 BIN DIF} (ix, jy) of the secondary binary television image of the detector for changing the background are formed from the signals of the low-pass filter S_ fil _DIF (ix, jу) in accordance with the rule:
L _{2 BIN} (ix, jy) = 1 if S_ fil _DIF (ix, jy)> Porog _DIF 1 and L 1 _{BIN DIF} (ix, jy) = 1
or S_ fil _DIF (ix, jy)> Porog _DIF 0 and L _{1 BIN DIF} (ix, jy) = 0,
otherwise L _{2 BIF DIF} (ix, jy) = 0,
where Rorog _DIF 1, Rorog _DIF 0 are the decision thresholds for the unit and zero elements of the primary binary television image of the background change detector, respectively. 36. The method according to p. 34, characterized in that the signals of the secondary binary television image L _{2 BIN SDO} (ix, jy) of the selector of moving objects are formed from signals S_ fil _CDO (ix, jy) of the low-pass filter in accordance with the rule:
L _{2 BIN DLS} (ix, jy) = 1 if S_ fil _LMS (ix, jy)> Porog _SDO 1 and L 1 _{BIN DLS} (ix, jy) = 1
or S_ fil _SDO (ix, jy)> Porog _CDO 0 and L _{1 BIN SDO} (ix, jy) = 0,
otherwise L _{2 BIN LMS} (ix, jy) = 0,
where Porog _SDO 1, Porog _SDO 0 - values of decision thresholds for single and zero elements of the primary binary television image of the selector of moving objects, respectively. 37. The method according to p. 36, characterized in that the horizontal and vertical projection G _{pr SDO} (ix), V _{pr SDO} (jy) signals of the secondary binary television image L _{2BIN SDO} (ix, jy) of the moving object selector is determined in accordance with

for ix = 1,. . . , Nwin,

for jy = 1,. . . , Mwin,
where Nwin and Mwin are the sizes of the secondary binary television image of the selector of moving objects horizontally and vertically, respectively. 38. The method of claim. 35 wherein the horizontal and vertical _straight projection _DIF G (ix), V _{ave DIF} (jy) of the secondary binary signal television image L _{2BIH DIF} (ix, jy) detector background changes determined in accordance with

for ix = 1,. . . , Nwin,

for jу = 1,. . . , Mwin,
where Nwin and Mwin are the sizes of the secondary binary television image of the background and horizontal changes detector, respectively. 39. The method according to p. 29, characterized in that the horizontal and vertical projections G _{pr GK} (ix), V _{pr GK} (jy) signals of the binary television image L _{BIN GK} (ix, jy) of the histogram classifier is determined in accordance with

for ix = 1,. . . , Nwin,

for jу = 1,. . . , Mwin,
where Nwin and Mwin are the dimensions of the binary television image of the histogram classifier horizontally and vertically, respectively. 40. The method according to p. 29, characterized in that the confidence coefficients W _DIF , W _LMS , W _{GK are} defined as the product of the input functions of the initial conditions and the normalized average densities of binary television images of the background change detector, moving object selector and histogram classifier, respectively

moreover, the averaged densities

binary television images of the change detector for handicap, selector of moving objects and a histogram classifier are obtained by limiting the minimum and maximum values and subsequent averaging by recursive first-order filters of the current densities V _DIF (n), V _SDO (n), V _GK (n) of the corresponding binary television images
Where

S _ODIF (n), S _OSDO (n), S _OGK (n) - current areas of binary television images of the background change detector, moving object selector and histogram classifier, respectively, inside the boundaries of the television image of the object;
S _OO (n) is the current area of the area inside the boundaries of the television image of the object;
F _{NU_DIF} (n), F _{NU_SDO} (n), F _{NU_GK} (n) - the input function of the initial conditions of the detector of changes in the background, the selector of moving objects and the histogram classifier;
n is the number of the current frame. 41. The method according to PP. 37,38,39 and 40, characterized in that the generalized horizontal and vertical _straight projection G _O (ix, n), V _{prosp O} (jy, n) signal of television generalized binary object image is formed of binary weighted summation projection television pictures detector background changes , moving object selector and histogram classifier:
G _prO (ix, n) = W _DIF (n) * G _prDIF (ix) + W _SDO (n) * G _{pr SDO} (ih) + W _GK (n) * G _pr GK (ih),
V _prO (jy, n) = W _DIF (n) * V _prDIF (jy) + W _SDO (n) * V _{pr SDO} (jу) + W _GK (n) * V _{pr GK} (jy),
where n is the number of the current frame. 42. The method according to p. 41, characterized in that the current coordinates X _OBIN , Y _{OBIN of the} generalized binary television image of the object is defined as the weighted sum of the coordinates of the center of gravity X _OCT , Y _OCT and the coordinates X _OMED , Y _{OMED of the} median area of the generalized binary television image of the object
X _OBIN (n) = W _CT (n) * X _OCT + [1-W _CT (n)] * X _OMED ,
Y _OBIN (n) = W _CT (n) * Y _OCT + [1-W _CT (n)] * Y _OMED ,
moreover, the weight coefficient W _CT (n) of estimating the coordinates of the center of gravity of the generalized binary television image of the object is increased with decreasing average deviation of the coordinates of the television image of the object from their predicted values. 43. The method according to p. 29, characterized in that the current horizontal VG _{OB_BIN} and vertical VV _{OB_BIN} components of the estimation of the speed of the generalized binary television image of the object in the inertial coordinate system are determined in accordance with the expressions
VG _{OB_BIN} = (dX + ΔX _OA + ΔX _{OB_OA_BIN} ) / T,
VV _{OB_BIN} = (dY + ΔY _OA + ΔY _{OB_OA_BIN} ) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the monitoring system during time T between obtaining the current and previous fields of the television image horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OB_OA_BIN} , ΔY _{OB_OA_BIN} - change the coordinates of the generalized binary television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively. 44. The method according to p. 29, characterized in that the signals of the dynamic reference television image of the object are formed by reading in each frame the signals from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET Whose} Y _{ET ET} is determined by the difference
X _{C ET} = X _{AB PZ} -X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA PZ} ,
wherein X _{ON PP,} Y _{ON PP} - the coordinates of the television picture of the object in the surveillance video camera system view;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system. 45. The method according to p. 29, characterized in that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , Y _{C ET} which is determined by the difference
X _{C ET} = X _{AB PZ} - X _{OA PZ} ,
Y _{ET C} = Y _{ON PP} - _PP Y _OA,
wherein X _{ON PP,} Y _{ON PP} - the coordinates of the television picture of the object in the surveillance video camera system view;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system;
when the conditions for changing the static reference television image of the object formed by comparing the parameters of the signals of the measure of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static and dynamic reference television images of the object, as well as comparing the parameters of the trajectories of the television image of the object obtained using signals static and dynamic reference television images of the object. 46. The method according to p. 29, characterized in that the signals of the static reference television image of the object are formed by reading and storing the signals of the television image from the current analysis window in a rectangular window with dimensions equal to the dimensions of the generalized binary television image of the object, with the center, coordinates X _{C ET} , Y _{C ET} which is determined by the difference
X _{C ET} = X _{AB PZ} -X _{OA PZ} ,
Y _{C ET} = Y _{ABOUT PZ} -Y _{OA PZ} ,
wherein X _{ON PP,} Y _{ON PP} - the coordinates of the television picture of the object in the surveillance video camera system view;
X _{OA PZ} , Y _{OA PZ} - coordinates of the analysis window in the field of view of the video camera of the surveillance system;
when the conditions for changing the static reference television image of the object, formed on the basis of the analysis of the signal parameters of the measure of dissimilarity of the signals of the television image after non-linear high-pass filtering in the current analysis window and the signals of the static reference television image of the object, as well as the path parameters of the television image of the object obtained on the basis of signal analysis measures of dissimilarity of television images. 47. The method of claim. 29, characterized in that the current horizontal VG _{_NSKH ON} and _ON VV _{_NSKH} vertical components of the rate of movement of a television image of an object in an inertial coordinate system obtained on the basis of the signal generating dissimilarity measure television images, is determined in accordance with expressions
VG _{OB_NXX} = (dX + ΔX _OA + ΔX _{OB_OA_NXX} ) / T,
VV _{OB_NX} = (dY + ΔY _OA + ΔY _{OB_OA_NX} ) / T,
where dX, dY is the movement of the axis of the field of view of the video camera of the surveillance system over time;
T between obtaining the current and previous fields of the television image horizontally and vertically, respectively;
ΔX _OA , ΔY _OA - the change in the position of the analysis window in the current frame relative to the previous frame horizontally and vertically, respectively;
ΔX _{OA_OA_NXX} , ΔY _{OA_OA_NXX} - change the coordinates of the television image of the object in the analysis window in the current frame relative to the previous frame horizontally and vertically, obtained on the basis of the formation of signals of the measure of dissimilarity of television images. 48. The method according to p. 29, characterized in that the confidence factor W _BIN (n) of the current velocity of the generalized binary television image of the object is obtained by determining the current density

generalized binary television image, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image, then averaging the limited density of the generalized binary television image with a first order recursive filter, normalizing the average density

generalized binary television image of an object

where S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object;
S _OO (n) is the current area of the region inside the boundaries of the generalized binary television image of the object;
n is the number of the current frame;

- average similarity coefficient obtained by determining the current similarity coefficient

restrictions on its maximum and minimum values and averaging by a first order recursive filter;
σ _{f min} (n) is the minimum rms value of the background television image signals in the background analysis windows;
Е _{НСХ min} (n) is the minimum value of the signals of the measure of dissimilarity of the television image in the current analysis window and the static reference television image of the object in the two-dimensional region of the search for the displacements of the television image of the object. 49. The method according to p. 29, characterized in that the confidence coefficient W _CH (n) of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of the television images, is obtained by determining the current coefficient of similarity

limiting its maximum and minimum values and averaging with a first-order recursive filter, normalizing the obtained average similarity coefficient

where n is the number of the current frame;
σ _fmin (n) is the minimum rms value of the background television image signals in the background analysis windows;
E _{HCX min} (n) is the minimum value of the signals of the measure of dissimilarity of the television images of the current analysis window and the static reference television image of the object in the two-dimensional region of the search for the displacements of the television image of the object;

- the average density of the generalized binary television image of the object obtained by determining the current density

generalized binary television image of the object, limiting the minimum and maximum values of the current density V _BIN (n) of the generalized binary television image of the object and then averaging, with a first order recursive filter, the limited current density V _BIN (n) of the generalized binary television image of the object;
S _OBIN (n) is the current area of the generalized binary television image inside the boundaries of the generalized binary television image of the object,
S _ОО (n) is the current area of the region inside the boundaries of the generalized binary television image of the object. 50. The method according to p. 29, characterized in that a comprehensive assessment of the horizontal VG _OB (n) and vertical VV _OB (n) components of the current speed of the television image of the object in the inertial coordinate system is determined by limiting estimates of the speed of the binary television image of the object and speed movement of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, the minimum and maximum values generated taking into account the previous values of a complex estimate of the current speed of moving a television image of an object and forming a weighted sum of limited estimates of the speed of moving a binary television image of an object and the speed of moving a television image of an object, obtained on the basis of the formation of signals of a measure of dissimilarity of television images
VG _OB (n) = W _BIN (n) * VG _{OB_BIN} (n) + W _CX (n) * VG _{ABOUT NSC} (n),
VV _OB (n) = W _BIN (n) * VV _{OB_BIN} (n) + W _CX (n) * VV _{ABOUT NSC} (n),
where W _BIN (n), W _CX (n) are the reliability coefficients of the current velocity of the generalized binary television image of the object and the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images, respectively;
VG _{OB_BIN} (n) and VV _{OB_BIN} (n) - limited horizontal and vertical components of the current speed of movement of the generalized binary television image of the object;
VG _{ABOUT NSX} (n) and VV _{ABOUT HCX} (n) - limited horizontal and vertical components of the current speed of the television image of the object, obtained on the basis of the formation of signals of the measure of dissimilarity of television images;
n is the number of the current frame. 51. The method according to p. 29, characterized in that the analysis of the current S _OBIN (n) and averaged

the area of the generalized binary television image of the object is produced by checking the fulfillment of the conditions:

- to go to the formation of control signals for the movement of the axis of the field of view of the video camera of the surveillance system by extrapolated coordinates of the television image of the object or condition,

- to proceed to the formation of control signals for the movement of the axis of the field of view of the video camera of the monitoring system according to the coordinates of the television image of the object in the field of view of the video camera of the monitoring system obtained by processing the television image in the current analysis window, where ksl is a constant coefficient, ksl <l; ks2 (nn _ek ) - coefficient decreasing with increasing frame number n, starting from frame number n _ek , the transition to the formation of control signals for moving the axis of the camera’s field of view along extrapolated coordinates, ks2 (nn _ek ) ≤ksl. 52. A signal processing device for determining the coordinates of objects observed in a sequence of television images, comprising a processor (95) for computing local data in time division by a common bi-directional bus, characterized in that a block (75) for receiving and storing signals of the current field of the television images from the video camera of the system for monitoring and generating signals for synchronizing the operation of the device, a block (76) for receiving and storing signals of uncontrolled movement and roll of the field of view de-cameras of the surveillance system, block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between reception of the signals of the previous and current fields of the television image, imager (78) of the signals of the television image of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the field axis view of the video camera of the monitoring system during the time between receiving signals of the previous and current fields of the television image and signals of the current field of bodies Viziona image elementwise coordinate transformation signals of the current field of television images providing displacement compensation unmanaged
and the roll of the field of view of the video camera of the surveillance system, a shaper (79) of the television image signals in the current analysis window using the position and size signals of the current television image analysis window, shaper (80) of the background analysis windows around the perimeter of the analysis window and determining the projections of the binary television image signals of the histogram classifier in the background analysis windows, block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, the switch (84) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, an analyzer (85) of the current velocity of a generalized binary television image of an object in an inertial coordinate system, an analyzer (86) of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, block (87) for determining the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, block (88) for determining the current velocity of a generalized binary television image of an object in an inertial coordinate system, block (89) forming a comprehensive assessment of the current speed of the television image of the object in an inertial coordinate system, block (90) for an integrated estimation of the coordinates of a television image of an object in the field of view of a video camera of a surveillance system, an analyzer (91) of conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, block (92) averaging a comprehensive assessment of the current speed of the television image of the object and storing the values of the filtered speed, an extrapolator (93) of coordinates and speed of the television image of the object in the next frame based on the analysis of the values of the average speed of the television image of the object, shaper (94) of signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and the speed of the television image of the object, moreover, the first and second inputs of the device are connected to the first and second inputs of the block (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating synchronization signals of the device, respectively, the third input of the device is connected to the input of the block (76) for receiving and storing signals of uncontrolled movement and the roll of the field of view of the video camera of the surveillance system, the fourth and fifth inputs of the device are connected to the first and second inputs of the switch (84) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, respectively, the sixth input of the device is connected to the second input of the block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals from the previous and current fields of the television image, the first output of the unit (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device is connected to the first input of the imager (78) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view video cameras of the monitoring system during the time between reception of signals of the previous and current fields of the television image and signals of the current field of television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, the second output of the block (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device is connected to the third input of the block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between receiving signals from the previous and current fields television image the fifth input of the imager (79) of the television image signals in the current analysis window using the position and size signals of the current television image analysis window, the fourth input of the imager (78) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the signals of the previous and current fields of the television image and the signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current television image fields providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, the second inputs of the shaper (80) of the background analysis windows around the perimeter of the analysis window and determine the projections of the signals of the binary television image of the histogram classifier in the background analysis windows and the block (87) for determining the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the third input of the block (81) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, the fourth input of the block (83) determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, the fifth input of the switch (84) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, the fourth inputs of the analyzer (85) of the current velocity of the generalized binary television image of the object in an inertial coordinate system, and an analyzer (86) the current speed of the television image of the object in an inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the fourth input of the block (88) determining the current velocity of the generalized binary television image of the object in an inertial coordinate system, the fifth input of the block (89) forming a comprehensive assessment of the current speed of the television image of the object in an inertial coordinate system, the third input of the block (90) of an integrated assessment of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the sixth input of the analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the third inputs of the averaging unit (92) for an integrated estimate of the current speed of moving the television image of the object and storing the values of the filtered speed and extrapolator (93) coordinates and the speed of moving the television image of the object in the next frame based on the analysis of the values of the average speed of moving the television image of the object and the sixth input of the shaper ( 94) control signals for moving the axis of the field of view of the video camera of the surveillance system and signals for the position and size of the analysis window in blowing television picture frame using the coordinates of the object in the television picture field of view of video camera surveillance system or extrapolated coordinates and velocity of the television image of the object, the first output of the block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of signals from the previous and current fields of the television image is connected to the second input of the signal generator (78) of the television image of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis field of view of the video camera of the monitoring system during the time between the reception of signals of the previous and current fields of the television image and signals of the current fields of the television image with elementwise transformation of the coordinates of the signals of the current field of the television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, the output of the block (76) for receiving and storing signals of uncontrolled movement and the roll of the field of view of the video camera of the monitoring system is connected to the third input of the imager (78) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system for the time between receiving signals of the previous and current fields of the television image and signals of the current field of the television image with elementwise coordinate transformation tension current field signals television image, providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, with the fifth input of the block (81) determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the first input of the block (82) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, with the first input of the block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, with the second input of the block (88) for determining the current velocity of the generalized binary television image of the object in the inertial coordinate system and with the first input of the block (90) for the complex estimation of the coordinates of the television image of the object in the field of view of the video camera of the surveillance system, the output of the imager (78) of the television image signals of the current frame from the television image signals of the previous frame, taking into account the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between the reception of the signals of the previous and current fields of the television image and the signals of the current field of the television image with elementwise transformation of the coordinates of the signals of the current field television image providing compensation for uncontrolled movements and the roll of the field of view of the video camera of the surveillance system, connected to the first input of the imager (79) of the television image signals in the current analysis window using the position and size signals of the current television image analysis window, the output of the imager (79) of the television image signals in the current analysis window using the position and size signals of the current television image analysis window is connected to the seventh input of the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image object and with the third input of the block (82) determining the coordinates of the television image of the object relative to the center of a current analysis window based on signal generation measures of dissimilarity of television images, the output of the shaper (80) of the background analysis windows along the perimeter of the window for analyzing and projecting the signals of the binary television image of the histogram classifier in the background analysis windows is connected to the first input of the block (87) for determining the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the first output of which is connected to the third input of the analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the first output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the shaper (80) of the background analysis windows along the perimeter of the analysis window and determine the projections of the signals of the binary television image of the histogram classifier in the background analysis windows, the second output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the switch (84) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the surveillance system, the first input of the analyzer (85) of the current velocity of the generalized binary television image of the object in the inertial coordinate system and the first input of the shaper (94) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system or extrapolated coordinates and the speed of movement of the television and siderations object the third output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the block (88) for determining the current speed of the movement of the generalized binary television image of the object in the inertial coordinate system, the fourth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the second input of the analyzer (85) of the current speed of the movement of the generalized binary television image of the object in the inertial coordinate system and to the second input analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on those uschey and averaged area generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the fifth output of the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the first input of the analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image images of the object the current average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the sixth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the third input of the analyzer (86) of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the seventh output of the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the fourth input of the block (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating synchronization signals device operation the first output of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the second input of the block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the third input of the block (89) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the second output of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the first input of the analyzer (86) of the current speed of the television image of the object in the inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the first output of which is connected to the second input of the block (89) for forming a complex estimate of the current speed of movement of the television image of the object in an inertial coordinate system, the first output of the switch (84) of codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the first input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the seventh input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the third input of the shaper (79) of the signals of the television image in the current analysis window using the position and size signals of the current analysis window of the television image, the second output of the switch (84) of codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the monitoring system is connected to the second input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the sixth input of the unit (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images and the second input of the imager (79) of the signals of the television image in the current analysis window using the position and size signals of the current analysis window of the television image, the first output of the analyzer (85) of the current velocity of the generalized binary television image of the object in the inertial coordinate system is connected to the first input of the block (89) for generating a complex estimate of the current speed of the television image of the object in the inertial coordinate system, and the second output is connected to the second input of the analyzer (86) of the current speed of the television image of the object in an inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, the second output of which is connected to the third input of the analyzer (85) of the current velocity of the generalized binary television image of the object in an inertial coordinate system, the first output of the block (88) for determining the current moving speed of the generalized binary television image of the object in the inertial coordinate system is connected to the fourth input of the block (89) for generating a complex estimate of the current speed of moving the television image of the object in the inertial coordinate system, the first output of the unit (89) for generating a comprehensive estimate of the current speed of the television image of the object in the inertial coordinate system is connected to the fourth input of the analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, the fifth input of the block (82) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images, the first input of the unit (92) for averaging the complex estimate of the current speed of the television image of the object and storing the values of the filtered speed and the second input of the block (90) for the complex estimate of the coordinates of the television image of the object in the field of view of the video camera of the monitoring system, the first output of the unit (90) for the complex assessment of the coordinates of the television image of the object in the field of view of the video camera of the monitoring system is connected to the fourth input of the switch (84) codes of the source or current dimensions of the object and the coordinates of the object in the field of view of the video camera of the monitoring system, the first input of the extrapolator (93) coordinates and the speed of the television image of the object in the next frame based on the analysis of the values of the average speed of the television image of the object and the third input of the driver (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of a video camera of a surveillance system or ex rapolirovannyh coordinates and velocity of the television image of the object, the second output of which is connected to the third inputs of the block (83) for determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of the formation of the signals of the measure of dissimilarity of the television images and the block (88) for determining the current velocity of the generalized binary television image of the object in an inertial coordinate system, the sixth input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object and the fourth inputs of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of television dissimilarity of images and a shaper (79) of television image signals in the current analysis window using We receive signals of the position and size of the current analysis window of the television image the first output of the analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows is connected to the second input of the generator (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field view of the video camera of the surveillance system or extrapolated coordinates and the speed of movement of the television image of the object, the second input of the unit (92) for averaging a complex estimate of the current speed of moving the television image of the object and storing the filtered speed and with the fourth input of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the second output of the analyzer (91) of the conditions for using the extrapolated coordinates of the object using the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows is connected to the second input of the unit (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images, the fourth output of the analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows is connected to the third input of the block (75) for receiving and storing signals of the current field of the television image from the video camera of the monitoring system and generating signals for synchronizing the operation of the device, the first output of the averaging unit (92) for complex estimation of the current speed of moving the television image of the object and storing the values of the filtered speed is connected to the fifth input of the analyzer (91) of the conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows and the second input of the extrapolator (93) coordinates and speed of the television image of the object in the next frame based on the analysis of the values of the average speed of the television image of the object, the second output of the unit (92) for averaging a complex estimate of the current speed of moving the television image of the object and storing the values of the filtered speed is connected to the fifth input of the shaper (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image with using the coordinates of the television image of the object in the field of view of the video camera of the surveillance system or extrapolated coordinates and speed television image of the object the first output of the extrapolator (93) coordinates and the speed of the television image of the object in the next frame, based on the analysis of the values of the average speed of the television image of the object, is connected to the fourth input of the driver (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the observation system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera system is observed or extrapolated coordinates and speed of the television image of the object, the first output of the driver (94) of the control signals for moving the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of the video camera of the monitoring system or extrapolated coordinates and the speed of movement of the television image of the object is the output device and is connected to the first input of the block (77) for calculating the controlled movement of the axis of the field of view of the video camera system s observation of the time between the reception signals of the previous and current fields of the television image, the output of the processor (95) for computing local data in a time-sharing shared bi-directional bus connected to the second outputs of the block (77) for calculating the controlled movement of the axis of the field of view of the video camera of the monitoring system during the time between receiving signals from the previous and current fields of the television image, block (88) determining the current velocity of the generalized binary television image of the object in an inertial coordinate system, block (83) determining the current speed of the television image of the object in the inertial coordinate system using the coordinates of the television image of the object, obtained on the basis of signal generation measures of dissimilarity of television images, block (87) determining the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, block (89) forming a comprehensive assessment of the current speed of the television image of the object in an inertial coordinate system, unit (90) for an integrated assessment of the coordinates of a television image of an object in the field of view of a video camera of a surveillance system, extrapolator (93) coordinates and speed of the television image of the object in the next frame based on the analysis of the values of the average speed of the television image of the object, the third outputs of the analyzer (85) of the current velocity of the generalized binary television image of the object in an inertial coordinate system, analyzer (86) the current speed of the television image of the object in an inertial coordinate system, obtained on the basis of signal generation measures of dissimilarity of television images, an analyzer (91) of conditions for using the extrapolated coordinates of the television image of the object based on the current and average area of the generalized binary television image of the object, the current and average speed of the television image of the object, the area and coordinates of the boundaries of the binary television images of the histogram classifier in the background analysis windows, block (92) averaging a comprehensive estimate of the current speed of the television image of the object and storing the values of the filtered speed and the driver (94) of the signals for controlling the movement of the axis of the field of view of the video camera of the monitoring system and the position and size of the analysis window in the next frame of the television image using the coordinates of the television image of the object in the field of view of a video camera of a surveillance system or extrapolated coordinates and speed of movement of a television image facility the eighth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object.  53. The device according to p. 52, characterized in that the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object contains a video switch (31), the first buffer random access memory (30 ), a second buffer random access memory (41) and a third buffer random access memory (43), a shaper (32) of signals of a binary television image the classifier classifier in the current analysis window, the shaper (36) of the signals of the primary binary television image of the detector of changes in the background, the shaper (38) of signals of the secondary binary television image of the detector of changes in the background, the shaper (45) of signals of the primary binary television image of the selector of moving objects, the shaper (47) signals of the secondary binary television image of the selector of moving objects, the first node (42) scaling and shifting the signals of the television the image, the node (44) for generating the signals of the differential television image of the selector of moving objects, the shaper (33) of horizontal and vertical projections of the signals of the binary television image of the histogram classifier, the shaper (39) of horizontal and vertical projections of the signals of the secondary binary television image of the detector for background changes, the shaper (48 ) horizontal and vertical projections of the signals of the secondary binary television image of the selector of moving objects, bl to (35) determining the object / background ratio and the minimum rms value of the background television image signals in the background analysis windows, a spatial low-pass filter (37) of the primary binary television image signals of the background change detector, a spatial low-pass filter (46) of the primary binary television image signals moving object selector, analyzer (34) of signals of binary television images of a histogram classifier, analyzer (40) of signals of secondary binary television images of the background change detector, an analyzer (49) of signals of secondary binary television images of the selector of moving objects and a generator (50) of generalized horizontal and vertical projections of signals of the generalized binary television image of the object, a computer (51) of the coordinates and dimensions of the object, the current and average area of the generalized binary a television image of the object, and an analyzer (52) of the conditions for the breakdown of the automatic determination of the coordinates of the television image of the object, and , the first and second inputs of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object are connected to the first and second inputs of the signal generator of the binary television image of the histogram classifier in the current analysis window and unit (35) for determining the ratio of the object / background and the minimum root mean square value of the signals of the television background image in approx on the background analysis and the fourth and third inputs of the primary binary television image signal generator (36) of the background change detector, respectively, the first input of the unit (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is also connected with the first input of the first node (42) scaling and shifting the signals of the television image, the third input of the block (81) is determined The current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object are connected to the third input of the video data switch (31), the seventh input of the histogram binary television image signal (32) signal in the current analysis window, the second input of the generator (33) horizontal and vertical projections of signals of a binary television image of a histogram classification the second input of the analyzer (34) of signals of binary television images of the histogram classifier, the fifth input of the block (35) for determining the ratio of object / background and the minimum rms value of the signals of the television image in the background analysis windows, the sixth input of the imager (36) of the signals of the primary binary television image detector of changes in the background, the second input of the analyzer (40) of the signals of the secondary binary television images of the detector for changes in the background, the fourth input of the first node (42) scaling and shifting signals of the television image, the second input of the analyzer (49) of the signals of the secondary binary television images of the selector of moving objects, the seventh input of the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized binary television image of the object and the second input of the calculator (51) of the coordinates and dimensions of the object , the current and average area of the generalized binary television image of the object, the fourth input of the block (81) determine the current coordinates of using a generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object is connected to the fourth input of the histogram classifier binary signal image generator (32) in the current analysis window, the fifth input of the general coordinates of the binary television image of the object is determined (81) using generalized horizontal and vertical projections of a generalized binary television of the image of the object is connected to the first input of the generator of the primary binary television image signals of the background change detector and the second input of the first node (42) of scaling and shifting the signals of the television image, the sixth input of the block (81) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of a generalized binary television image of the object is connected to the second input of the shaper (36) of the signals the primary binary television image of the background change detector, the fifth input of the histogram classifier of the binary television image of the histogram classifier signals in the current analysis window and the third input of the block (35) for determining the object / background ratio and the minimum rms value of the background image signal in the background analysis windows, seventh block input (81) for determining the current coordinates of a generalized binary television image of an object using generalized horizontal and vertical of projections of the generalized binary television image of the object is connected to the inputs of the first buffer random access memory (30) and the second buffer random access memory (41), the first input of the differential TV image generating unit (44) of the signals of the moving objects selector, the fifth input of the signal shaper (36) the primary binary television image of the background change detector and the first input of the video data switch (31), the output of the first buffer memory device (30) is connected to the second input of the video data switch (31), the first and second outputs of which are connected to the third and sixth inputs of the signal generator of the binary television image of the histogram classifier in the current analysis window, respectively, and to the fourth input of the determination unit (35) the object / background ratio and the minimum rms value of the background television image signals in the background analysis windows, the first output of which is the fifth output of the current coordinate determination unit (81) generalized of a binary television image of an object using generalized horizontal and vertical projections of a generalized binary television image of an object, the first output of the histogram classifier binary signal image generator (32) in the current analysis window is connected to the first output of the generalized binary television image image determination unit (81) using generalized horizontal and vertical projections of a generalized binary television image of the object and with the first input of the shaper (33) of horizontal and vertical projections of the binary television image signals of the histogram classifier, the output of which is connected to the first input of the shaper (50) of generalized horizontal and vertical projections of the signals of the generalized binary television image of the object and with the first input of the analyzer (34) signals of binary television images of a histogram classifier, the first output of which is connected to the second input of the generator (50) of generalized the horizontal and vertical projections of the signals of the generalized binary television image of the object, the output of the shaper (36) of the signals of the primary binary television image of the background change detector is connected to the input of the spatial filter (37) of the low frequencies of the signals of the primary binary television image of the background change detector, the output of which is connected to the input of the shaper ( 38) signals of the secondary binary television image of the background change detector, the output of which is connected to the input of the Atelier (39) of horizontal and vertical projections of signals of the secondary binary television image of the background change detector, the output of which is connected to the third input of the generator (50) of generalized horizontal and vertical projections of signals of the generalized binary television image of the object and with the first input of the analyzer (40) of signals of secondary binary television images of the background change detector, the first output of which is connected to the fourth input of the shaper (50) of generalized horizontal and vertical projections of signals of the generalized binary television image of the object, the output of the second buffer random access memory (41) is connected to the third input of the first node (42) of scaling and shifting the signals of the television image, the output of which is connected to the input of the third buffer random access memory (43), the output of which is connected with the second input of the node (44) for generating signals of the differential television image of the selector of moving objects, the output of which is connected to the input of the shaper (45) signal of the primary binary television image of the moving object selector, the output of which is connected to the input of the spatial filter (46) of the low frequencies of the signals of the primary binary television image of the moving object selector, the output of which is connected to the input of the generator of the secondary binary television image signal (47) of the moving object, the output of which connected to the input of the shaper (48) of horizontal and vertical projections of the signals of the secondary binary television image the selector of moving objects, the output of which is connected to the first input of the analyzer (49) of the signals of the secondary binary television images of the selector of moving objects and the sixth input of the generator of (50) generalized horizontal and vertical projections of the signals of the generalized binary television image of the object, the first output of the analyzer (49) of signals of the secondary of binary television images of the moving object selector is connected to the fifth input of the shaper (50) of generalized horizontal and vertical projections of signals Generalized binary television image of the object, the first output of which is connected to the first input of the computer (51) of coordinates and dimensions of the object, the current and average area of the generalized binary television image of the object and with the fourth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using generalized horizontal and vertical projections of a generalized binary television image of an object, the first and second outputs of a computer (51) coordinate t and the size of the object, the current and average area of the generalized binary television image of the object are the second and third outputs of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, respectively, the third output of the calculator (51 ) coordinates and dimensions of the object, the current and average area of the generalized binary television image of the object with the input of the analyzer (52) of the conditions for the breakdown of the automatic determination of the coordinates of the television image of the object, the first output of which is the seventh output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the third output of the block (35) determining the object / background ratio and the minimum rms value of the background television image signals in The analysis of the background is connected to the sixth output of the block (81) for determining the current coordinates of the generalized binary television image of the object using the generalized horizontal and vertical projections of the generalized binary television image of the object, the eighth output of which is connected to the second outputs of the signal generator (32) of the binary television image of the histogram classifier in the current analysis window, block (35) for determining the object / background ratio and the minimum rms value of the signals television background images in the windows of the background analysis, the analyzer (34) of the signals of the binary television images of the histogram classifier, the analyzer (40) of the signals of the secondary binary television images of the background change detector, the analyzer (49) of the signals of the secondary binary television images of the selector of moving objects, the analyzer (52) of the conditions failure of automatic determination of the coordinates of the television image of the object and the shaper (50) of the generalized horizontal and vertical projections of the signals of the generalized about the binary television image of the object and the fourth output of the computer (51) of the coordinates and dimensions of the object, the current and average area of the generalized binary television image of the object.  54. The device according to p. 53, characterized in that the generator (32) of the signals of the binary television image of the histogram classifier in the current analysis window contains a calculator (67) normalized histograms of the distribution of the brightness of the signals of the television images of the object and background and the node (68) of the formation of the binary television image histogram classifier, and, the first, second, third, fourth and fifth inputs of the generator (32) of signals of the binary television image of the histogram classifier in the current not analyzed are connected to the first, second, third, fourth and fifth inputs of the calculator (67) of the normalized histograms of the distribution of the brightness of the signals of the television images of the object and background, respectively, the sixth input is connected to the second input of the node (68) of the formation of the binary television image of the histogram classifier, the seventh input of the shaper (32) signals of the binary television image of the histogram classifier in the current analysis window is connected to the sixth input of the calculator (67) normalized histograms of the distribution the luminance of the signal brightness of the television images of the object and background and the third input of the node (68) forming the binary television image of the histogram classifier, the first output of the calculator (67) normalized histograms of the distribution of the brightness of the signals of the television image of the object and background is connected to the first input of the node (68) of the formation of the binary television image histogram classifier, the output of which is the first output of the generator (32) of signals of the binary television image of the histogram assifikatora in the current window of analysis, the second output of the calculator (67) of the normalized histogram distribution of the brightness signals of television pictures and the background object is connected to the output of the second (32) binary signals television image histogram classifier in the current analysis window.  55. The device according to p. 53, characterized in that the shaper (36) of the signals of the primary binary television image of the detector for changing the background comprises a shaper (70) of signals of the television image of the reference background, the first (69) and the second (72) buffer random access memory, node (71) scaling and shifting the signals of the television image of the reference background, the node (73) for generating the signals of the differential television image of the background change detector and the node (74) of the primary binarization of the background change detector, the first and second outputs of the primary binary television image signal generator (36) of the background change detector are connected to the first and second inputs of the television image signal generator (70) of the reference background signal and the node (71) of scaling and shifting the television image signals of the background background, the third and fourth inputs - connected to the third and fourth inputs of the imager (70) of the television image signals of the reference background, respectively, and the fifth input is connected to the first input of the node (73) image the signals of the differential television image of the background change detector and the input of the first buffer random access memory (69), the output of which is connected to the fifth input of the signal generator of the television image signal (70) and the sixth input of the signal generator (36) of the primary binary television image of the background detector connected to the third input of the signal generation unit (73) of the difference television image of the background change detector and the seventh input of the signal shaper (70) a television image of a reference background, the output of which is connected to the third input of the scaling and shifting unit (71) of signals of a television image of a reference background, the output of which is connected to the input of a second buffer random access memory (72), the output of which is connected to the sixth input of the signal generator (70) of the television images of the reference background and the second input of the node (73) for generating the signals of the differential television image of the background change detector, the output of which is connected to the input of the node (74) primarily Background binarization detector changes, the output of which is connected to the output of the shaper (36) of the primary binary signals television image detector background changes.  56. The device according to p. 52, characterized in that the unit (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images contains a switch (53) of signals of a scaled image, a non-linear high-pass filter (54) , an analyzer (55) of conditions for updating a static reference television image of an object, an analyzer (61) of the type of sequences of minimum values of signals of a measure of dissimilarity of television images to Only rows and along the columns of the two-dimensional region for searching for displacements of the television image of the object, the first (56) and second (59) buffer random access memory devices, the shaper (57) of the signals of the static and dynamic reference television images of the object, the shaper (60) of signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for displacements of the television image of the object, the node (58) for scaling the signals of the television image, the selector (62) of the sequence of minimum values the signals of the measure of dissimilarity of television images, the approximator (63) of the sequence of the minimum values of the signals of the measure of dissimilarity of television images by a polynomial of the fourth degree, the coordinator (64) of the television image of the object in the analysis window according to the minimum position of the approximating polynomial, the coordinator (65) of the television image of the object in the analysis window by offset boundaries of the region of rapid growth of signal values of the measure of dissimilarity of television images relative to the center of the analysis window, coordinator (66) of television of the image of the object in the analysis window according to the coordinates of the center of the analysis window, and the first input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the first inputs of the analyzer (55) of the conditions for updating the static reference of a television image of an object and a shaper (60) of signals of a measure of dissimilarity between signals of television images in the two-dimensional domain of the image of the object, the second input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the first input of the switch (53) of the signals of the scaled image, the third input of the block (82) for determining the coordinates of the television image object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of television images is connected to the second input of the switch (53) signals of the scaled image and with the first input of the nonlinear high-pass filter (54), the output of which is connected to the third input of the switch (53) of the signals of the scaled image, the fourth input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on of generating signals of the measure of dissimilarity of television images is connected to the third input of the generator (60) of signals of the measure of dissimilarity between signals of television images in the two-dimensional region of the search of the television image of the object and with the second inputs of the shaper (57) of the signals of the static and dynamic reference television images of the object and the analyzer (55) of the conditions for updating the static reference television image of the object, the fifth input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window on the basis of the formation of signals of the measure of dissimilarity of television images is connected to the fourth input of the generator (60) of signals of the measure of dissimilarity between by the television images in the two-dimensional region for searching for the displacements of the television image of the object, the sixth input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the generation of signals of the measure of dissimilarity of the television images is connected to the third input of the generator of the static and dynamic reference television signals (57) images of the object, the seventh input of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis windows based on the generation of signals of the measure of dissimilarity of television images is connected to the fourth input of the shaper (57) of the signals of the static and dynamic reference television images of the object and to the fifth input of the shaper (60) of signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region for searching for displacements of the television image of the object, the eighth input of the block (82) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation the signals of the measure of dissimilarity of television images is connected to the second input of the nonlinear high-pass filter (54), to the fifth input of the analyzer (55) of the conditions for updating the static reference television image of the object, the fourth input of the selector sequence (62) of the minimum values of the signals of the measure of dissimilarity of television images and the seventh input of the shaper (60) signals of the measure of dissimilarity between the signals of the television images in the two-dimensional region of the search for the displacements of the television image of the object, the output of the comm the scaler image signal generator (53) is connected to the second input of the signal shaper (60) of the dissimilarity measure between the television image signals in the two-dimensional region for searching for the television image offsets of the object and the input of the first buffer random access memory (56), the output of which is connected to the first input of the shaper (57 ) signals of static and dynamic reference television images of the object, the first output of the analyzer (55) update conditions of the static reference television image of the object is connected to the fifth input of the shaper (57) of signals of static and dynamic reference television images of the object, the output of which is connected to the input of the node (58) for scaling the signals of the television image, the output of which is connected to the input of the second buffer random access memory (59), the output of which is connected with the sixth input of the signal shaper (60) of the measure of dissimilarity between the signals of the television images in the two-dimensional region of the search for the displacements of the television image of the object, the output connected to the input of the analyzer (61) of the type of sequences of minimum values of signals of a measure of dissimilarity of television images along the rows and along the columns of the two-dimensional region for searching for the displacements of the television image of the object, the first output of which is connected to the second output of the block (82) for determining the coordinates of the television image of the object relative to the center of the current window analysis based on the formation of signals of a measure of dissimilarity of television images, the third input of the analyzer (55) conditions for updating this static of a linear television image of the object and the input of the selector (62) of the sequence of minimum values of the signals of the measure of dissimilarity of television images, the first output of which is connected to the input of the approximator (63) of the sequence of the minimum values of signals of the measure of the dissimilarity of television images of the fourth degree polynomial, the first output of which is connected to the input of the coordinator (64 ) a television image of the object in the analysis window by the position of the minimum of the approximating polynomial, the second and third outputs of the selector (62) last the sequence of the minimum values of the signals of the measure of dissimilarity of the television images are connected to the inputs of the coordinator (65) of the television image of the object in the analysis window by shifting the border of the region of rapid growth of the values of the signals of the measure of the dissimilarity of the television images relative to the center of the analysis window and the coordinator (66) of the television image of the object in the analysis window the center of the analysis window, the first outputs of the coordinator (64) of the television image of the object in the analysis window by the position of the approximating minimum about a polynomial, the coordinator (65) of the television image of the object in the analysis window by shifting the boundary of the region of rapid growth of signal values of the measure of dissimilarity of the television images relative to the center of the analysis window and the coordinator (66) of the television image of the object in the analysis window by the coordinates of the center of the analysis window are connected to the first output of the block (82) determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of a measure of dissimilarity of television images and The fourth input of the analyzer (55) for updating the static reference television image of the object, the third output of the block (82) for determining the coordinates of the television image of the object relative to the center of the current analysis window based on the formation of signals of the measure of dissimilarity of the television images is connected to the second outputs of the analyzer (61) of the type of sequences of minimum values signals of the measure of dissimilarity of television images along the rows and along the columns of the two-dimensional region of the search for the displacements of the television image the object, the analyzer (55) of the conditions for updating the static reference television image of the object, the approximator (63) of the sequence of the minimum values of the signals of the measure of dissimilarity of the television images by a polynomial of the fourth degree, the coordinator (64) of the television image of the object in the analysis window according to the minimum position of the approximating polynomial, coordinator (65 ) a television image of an object in the analysis window for the shift of the boundary of the region of rapid growth of signal values of the measure of dissimilarity of television images tnositelno analysis window center and focal point (66) the television image of the object in the analysis window center coordinates of the analysis window.

Images