JP3487045B2 - Automatic tracking device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic tracking device that uses an image to be picked up so as to automatically track a designated object in the center of the image.

[0002]

2. Description of the Related Art In recent years, automatic tracking devices that utilize images have come to be used in many fields because of their convenience. A template matching method has been conventionally used as an automatic tracking method using this image. The template matching method is a method in which image data cut out into a predetermined size called a template is moved on an input image, and a correlation calculation is performed at each moved position to find a position having the highest correlation.

By the way, in automatic tracking, first, a tracking target is specified on an image by using a target designation frame or the like, and image data in the target designation frame is stored as a template. Next, automatic matching is realized by applying template matching to the images that are input one after another and finding the position with the highest correlation.

In this template matching method, however, the shape and size of the tracking target on the image change due to the movement of the tracking target, etc., so that the correlation between the template and the input image collapses and tracking becomes impossible. There is. Therefore, in order to maintain the correlation, a method of updating the template every moment has been proposed. Further, as a means for further improving the tracking performance, a method of mounting a zoom lens mechanism and controlling the zoom lens mechanism based on the size of the extracted tracking target has been proposed.

Further, in the template matching method, the correlation calculation is performed while moving the template at a predetermined search step interval in the search area, so that the calculation amount is very large and the processing is time consuming and the tracking speed is hindered. There are cases. Therefore, in order to reduce the amount of calculation, a method of performing correlation calculation using all the image data in the template instead of thinning it out and performing a calculation process, a method of roughening the search step interval, and the like have been proposed.

On the other hand, the range in which the tracking target is searched on the image, that is, the search area, is an appropriate range if the positional relationship between the image input means and the tracking target, the motion model of the tracking target, and the image input interval can be defined. The tracking performance can be improved and the processing time can be shortened by limiting the search area. However, since the positional relationship between the tracking target and the image input means is usually not easily obtained, a method of using a fixed search area that is set wide considering the worst case, or simply on the image of the tracking target A method has also been proposed in which the motion of is expressed as a two-dimensional model and the position of the tracking target that appears next in the image is predicted to limit the search area.

[0007]

In the method of successively updating the template described above, for example, FIG.
As shown in (a), when the size of the tracking target P is smaller than that of the tracking gate G, the influence of the background is large. In such a case, the correlation with the background is more dominant than the tracking target, and the tracking target P is shown in FIG.
The tracking target P may not be tracked even if it moves as shown in (b). That is, in FIG. 10 (b), the result of the correlation operation, it towards the correlation tracking gate G ₂ position than the correlation tracking gate G ₁ position is high is desired, better correlation tracking gate G ₁ position There is a problem that it may become large.

In the method of controlling the zoom lens mechanism, since the size of the tracking target imaged by the image input means can be enlarged so as to always match the size of the tracking target of the template, the size of the tracking target is larger than that of the tracking target. Although the correlation can be prevented from becoming dominant, there is a problem that it is heavy and bulky and the cost increases.

The method of thinning out the image data in the template used for the correlation calculation at a constant rate is such that when the distance between the tracking target and the image input means is long, the tracking target appears small on the image, so that the tracking performance is reduced when thinning out. descend. Further, if the search area size is fixed, if the distance between the tracking target and the image input means is long, an unnecessary portion will be searched, and the tracking performance will deteriorate.

The use of the predicted position by the two-dimensional model of the tracking target is effective when the tracking target performs a constant velocity motion, but when the constant velocity motion is not performed, the prediction does not always hit and a necessary region is searched. There is a problem that you cannot do it.

Although the method of narrowing the search step interval reduces the amount of calculation, it has a problem of increasing the error in the position of the tracking target when the distance between the tracking target and the image input means is long.

The present invention has been made to solve the above problems, and an object thereof is to provide an automatic tracking device having excellent tracking performance.

[0013]

In order to solve the above problems, the present invention provides an image input means and a tracking target designating means for designating a tracking target in an image. Depending on the position of the tracking target calculated by the target position detecting unit and the position of the tracking target by moving the tracking gate within the specified search region in the image at a predetermined search step interval, An automatic tracking device including a rotating means for rotating the image input means, wherein the tracking gate size is changed according to a rotating position in the tilt direction,
A tracking gate size control means is provided.

According to the second aspect of the invention, the image input means, the tracking target designating means for designating the tracking target in the image, and the predetermined search step for the tracking gate in the designated search area in the image. An automatic tracking device comprising: a target position detection unit that moves at intervals to calculate the position of a tracking target; and a rotating unit that rotates an image input unit according to the position of the tracking target calculated by the target position detection unit. A search area size control means for changing the size of the search area in accordance with the rotational position in the tilt direction is provided.

[0015] In the invention described in claim 1, wherein, further
In addition, a thinning-out control means for changing the thinning-out rate of the image data used to calculate the tracking target position is added according to the rotational position in the tilt direction.

[0016] In the invention described in claim 2, further
In addition, a search step interval control means for changing the search step interval for searching the specified search area in the image according to the tilt position is added.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION A first embodiment of an automatic tracking device according to the present invention will be described below in detail with reference to FIGS. 1 to 7, and a second embodiment will be described in detail with reference to FIG. To do.

[First Embodiment] FIG. 1 is a block diagram showing an automatic tracking device, FIG. 2 is an image diagram for explaining a method of specifying a tracking target, and FIG. 3 is an image for explaining the operation of a tracking gate size control means. 4, FIG. 4 is a side view for explaining the relationship between the tracking target and the image input means, FIG. 5 is an image view for explaining the operation of the tracking gate size control means, and FIG. 6 is an image view for explaining the operation of the thinning control means. FIG. 7 is an image diagram for explaining the operation of the search means.

As shown in FIG. 1, the automatic tracking device includes an image input means 1, a rotating means 2, a tilt angle input means 3, and
The tracking target designating unit 4, the image monitor device 5, and the image processing unit 6 are included. The image input means 1 is a CCD
It is composed of a camera. The rotating means 2 is a turntable that supports the image inputting means 1, and changes the image pickup direction of the image inputting means 1 with two axes of the pan direction and the tilt direction. The tilt angle input means 3 is a potentiometer or the like attached to the tilt rotation shaft of the rotation means 2, and outputs a voltage signal proportional to the tilt angle θy of the image input means 1. The tracking target designating unit 4 is a console operated by an operator at the time of initial setting. The image monitor device 5 is composed of a CRT or the like, and displays an image input by the image input means 1 and an image processed by the image processing unit 6.

The image processing section 6 includes an image storage means 6a, a tracking gate position storage means 6b, a target storage means 6c, a search means 6d, a correlation calculation means 6e, a target position detection means 6f, and a rotation angle. The calculation means 6g, the tracking gate size control means 6h, and the thinning control means 6i are provided.

The image storage means 6a is composed of an image memory or the like, and stores the image data input from the image input means 1. The tracking gate position storage means 6b stores the current position of the tracking gate. The target storage means 6c stores image data to be used as a template. The search means 6d sequentially moves the tracking gate to a position where correlation calculation is performed within a predetermined search area and at predetermined search step intervals, and obtains image data required at each moving position from the image storage means 6a. Output to 6e.

The correlation calculation means 6e performs a correlation calculation based on the image data output from the image storage means 6a by the search means 6d and the image data stored in the target storage means 6c. The target position detecting means 6f obtains the position having the highest correlation based on the respective results calculated by the correlation calculating means 6e. The rotation angle calculation means 6g calculates a rotation angle between the pan direction and the tilt direction for giving an instruction to the rotation means 2, which is necessary for adjusting the tracking target position indicated by the coordinates on the image to the image center. To do. The tracking gate size control means 6h is
A signal proportional to the tilt angle obtained from the tilt angle input means 3 is converted into a tracking gate size. Thinning control means 6i
On the basis of the tracking gate size obtained by the tracking gate size control means 6h, thins out the image data used for the correlation calculation.

The above-mentioned automatic tracking device operates as follows. That is, the operator designates a tracking target as an initial setting before starting the automatic tracking. The image captured by the image input means 1 can be confirmed through the display screen 51 of the image monitor device 5, as shown in FIG. 2, for example. The display screen 51 displays a rectangular tracking gate G indicating a region of a template for storing a tracking target. Therefore, the operator uses the tracking target designating means 4 to
The rotating means 2 is operated so that the tracking target P enters the tracking gate G, and automatic tracking is started when the tracking target P enters the tracking gate G.

The tracking gate G is arranged so that the center O of the image is the center, as shown in FIG. Moreover, the size of the tracking gate G is variable so that the tracking target can be appropriately surrounded on the image. The size of the tracking gate G is set as follows. That is, for example, as shown in FIG. 4, the installation height of the image input means 1 is Hc, and the tracking target P is
Is Hp, and the horizontal distance between the image input means 1 and the tracking target P is Lp. The image captured by the image input means 1 shown in FIG. 4 is displayed on the display screen 51 of the image monitor device 5.
It is assumed that the image is displayed as shown in FIG.

At this time, the vertical size Sy of the tracking target P on the image displayed on the display screen 51 can be obtained by geometrical calculation and can be expressed by the equation (1).

Sy = F (f, Hc, Hp, θy) (1) That is, the vertical size Sy measures the installation height Hc of the image input means 1 in advance. Together with the height H of the tracking target P
If p is set, the focal length f of the image pickup system of the image input means 1
Is known, and is represented as a function of the tilt angle θy of the image input means 1. Further, the size S of the tracking target P in the horizontal direction is
x is a rough formula and is practically sufficient if it is expressed by the equation (2).
In the equation (2), C is a proportional constant.

Sx = C × Sy ………………………………………… (2) In this way, by changing the tracking gate size according to the tilt angle θy, the tracking gate It is possible to properly surround the tracking target. Further, the tilt angle θy can be converted into the tracking gate size by obtaining these relationships in advance and forming a table without depending on the arithmetic expression. The tracking gate size control means 6h is composed of this conversion table,
The horizontal size Sx and the vertical size Sy are output according to a signal proportional to the tilt angle θy output from the tilt angle input means 3.

When the tracking target designation operation is performed, the automatic tracking device stores the position of the tracking gate in the tracking gate position storage means 6b. Further, the automatic tracking device reads the image data in the tracking gate at the position set in the tracking gate position storage means 6b from the image storage means 6a and stores it in the target storage means 6c.

When the initial setting is completed as described above,
The automatic tracking device starts tracking. First, the image input means 1
Captures a predetermined image capturing area at regular time intervals and outputs the captured image as image data to the image storage means 6a. The image data output from the image input means 1 is the image storage means 6a.
Are sequentially stored in. The search means 6d reads the image data of the image storage means 6a corresponding to the tracking gate and outputs it to the correlation calculation means 6e. Upon completion of this reading, the searching means 6d moves the tracking gate within a predetermined search area by a predetermined search step interval.

The thinning-out control means 6i increases thinning-out when the tracking gate size is large, and does not perform thinning-out when the tracking gate size is small. This is because when the tracking gate size is large, the amount of correlation calculation is large and the processing time is long. On the other hand, when the tracking gate size is small, the resolution is low, so accurate tracking cannot be expected if thinning is performed. . Decimation is performed as follows.

The number of pixels for which the correlation calculation is performed is predetermined as Nx dots in the horizontal direction and Ny dots in the vertical direction. Then, if the tracking gate size output from the tracking gate size control means 6h is Sx × Sy dots, the pixel to be subjected to the correlation calculation uses the relative address coordinates (x, y) and the following (3) ( It can be expressed as in equation 4).

X = i × (Sx −1) / (Nx −1) …………………… (3) y = j × (Sy −1) / (Ny −1) …………… (4) However, in the equations (3) and (4), i is i = 0, 1, ...
(Nx -1) is a positive integer, and j is j = 0, 1, ...
It is a positive integer of (Ny -1), and x and y are positive integers with the fractional part truncated.

When the tracking gate size output from the tracking gate size control means 6h is input, the thinning-out control means 6i performs a correlation calculation so as to perform a correlation calculation only for pixels which are not thinned out according to the equations (3) and (4). Control means 6e. For example, the operation of the thinning control means 6i will be described below with reference to FIG. That is, the number of pixels for which a predetermined correlation calculation is performed is Nx = 6 dots in the horizontal direction and Ny = 14 dots in the vertical direction, and the tracking gate size output from the tracking gate size control means 6h is Sx = 8 dots in the horizontal direction. Assuming that Sy = 17 dots in the vertical direction, the value of x in equation (3) is x = 0,1,2,4,5,6, and the value of y in equation (4) is y = 0,1, 2,3,5,6,7,8,
It becomes 10,11,12,13,15,16.

That is, the tracking gate size control means 6h
Even if the tracking gate size output by the horizontal direction is Sx = 8 dots in the horizontal direction and Sy = 17 dots in the vertical direction, the pixels actually used for the correlation calculation are x = 0,1,2,4,5, 6 columns of 6 and y = 0,1,2,3,5,6,7,8,10,11,12,13,15,16 14 rows total 6 × 14 = 84 pixels, the rest The total of (8 × 17) −84 = 52 pixels of 2 columns of x = 3,7 and 3 rows of y = 4,9,14 is not used for the correlation calculation. FIG. 6 shows an image of the above-described thinning-out, in which hatched pixels are thinned out. At the time of the correlation calculation, the image data stored as the template in the target storage means 6c is also thinned based on the equations (3) and (4).

When the tracking gate size output from the tracking gate size control means 6h is less than a predetermined horizontal Nx dot and vertical Ny dot, the thinning control means 6i selects the same pixel rather than thinning. Pixels of horizontal direction Nx dots and vertical direction Ny dots that have been substantially expanded are used for the correlation calculation.

The correlation calculation means 6e is the thinning control means 6i.
The calculation based on the following equation (5) is performed based on the instruction from.

Σ (i, j) = Σ ₀ ^Nx Σ ₀ ^{Ny │Q} (x, y) -q (x, y) │ ... (5) That is, the correlation calculating means 6e is used for each corresponding image. The sum σ (i, j) of the absolute values of the differences between the pixel values in the tracking gate in the storage means 6a and the pixel values in the target storage means 6c (template) is obtained. That is, the sum σ (i, j) is a measure of inconsistency. Therefore, the image in the tracking gate in the image storage means 6a and the image in the target storage means 6c (template) are more similar as the total sum σ (i, j) is smaller, and the correlation is higher. Become. As the pixel value, a brightness value or a hue value is used.

In equation (5), (i, j) of σ (i, j)
j) represents the following. That is, the searching means 6d sequentially moves the tracking gate at a predetermined search step interval within a predetermined search area, and (5) for each moving position.
The equation is calculated, and (i, j) is an identifier attached corresponding to the position of each tracking gate at this time and represents a relative position. That is, as shown in FIG. 7, the tracking gate G moves δx dots per step in the horizontal direction and δy dots per step in the vertical direction. The tracking gate G moves I steps in the horizontal direction and J steps in the vertical direction, and ends the search over the entire area within the predetermined search area R. That is, the calculation result of the equation (5) regarding the tracking gate G (i, j) at a certain position is σ (i, j). Further, in the equation (5), (x, y) is a relative address coordinate determined based on the equations (3) and (4), and is a relative address coordinate of each pixel which is not the target of thinning when performing the correlation calculation. Represents Q (x, y) represents the pixel value of the relative address coordinate (x, y) in the image storage means 6a. q (x, y) represents the pixel value of the relative address coordinate (x, y) in the target storage means 6c (template).

The target position detecting means 6f calculates the position of the tracking target as follows. That is, the target position detecting means 6f has the smallest value σ (i) among the I × J σ (i, j) obtained as a result of the search over the entire area within the predetermined search area R. , j), and the position of the tracking target on the image stored in the image storage means 6a is obtained based on (i, j).

The rotation angle calculation means 6g calculates the pan direction rotation angle θx (6) necessary for aligning the position (i, j) having the highest correlation calculated by the target position detection means 6f with the image center.
The tilt direction rotation angle θy is calculated based on the equation (7) and is output to the rotating means 2. At this time, the target storage means 6c determines the tracking gate size of the image data input to the image storage means 6a by the image input means 1 according to the input signal from the tilt angle input means 3 by the tracking gate size control means 6h. Then, it is newly updated and stored.

[0041]         θx = tan^-1(L_x/ F) ………………………………………… (6)         θy = tan^-1(L_y/ F) ………………………………………… (7) In the equations (6) and (7), f is the image pickup system of the image input means 1.
Is the focal length of. L _xIs the target position detection means 6f
Center position of tracking gate G (i, j) that minimizes the target σ (i, j)
Image pickup of the position on the image pickup surface of the image input means 1 corresponding to
It is the horizontal distance from the plane center. L_yIs the target position
Tracking gate G that minimizes σ (i, j) obtained by the output means 6f
On the imaging plane of the image input means 1 corresponding to the center position of (i, j)
Is the distance in the vertical direction from the center of the imaging surface. Na
Oh, L_xOr L_yThe unit of is millimeter,
Horizontal length and vertical length per pixel of the image pickup surface
Are respectively known, and the horizontal deviation from the center of the imaging plane
If the number of dots for each of
_xOr L_yCan be easily calculated by proportional calculation.

The rotating means 2 is composed of the rotating angle calculating means 6g,
When a signal including the pan direction rotation angle θx and the tilt direction rotation angle θy is received, the image input means 1 is based on the signals.
The image pickup direction of is tracked and the tracking target P is rotated in the image pickup direction in which it is captured in the center of the image. Then, thereafter, the automatic tracking device realizes automatic tracking by repeating the above processing based on the image data newly stored in the target storage means 6c (template).

Therefore, in the automatic tracking apparatus of the first embodiment, the tracking gate size is changed according to the tilt angle θy of the image input means 1, so that the tracking target P on the image is the tracking gate. Since the area is always properly surrounded and the background portion of the image is not so much affected, the tracking performance can be improved. Further, since the thinning control is performed according to the tilt angle θy, even if the tracking target is located near the image input unit 1 and the tracking gate size increases and the number of pixels increases, the thinning is performed and the correlation calculation is performed. Since it is performed, the processing speed does not decrease.

[Second Embodiment] FIG. 8 is a block diagram showing an automatic tracking device, and FIG. 9 is an explanatory diagram showing the operation of the search area size control means. In FIG. 8, the same parts as those in the automatic tracking device described in the first embodiment are designated by the same reference numerals.

As shown in FIG. 8, the automatic tracking device includes an image input means 1, a rotating means 2, a tilt angle input means 3, and
The tracking target designating unit 4, the image monitor device 5, and the image processing unit 6 are included. The image input means 1 is a CCD
It is composed of a camera. The rotating means 2 is a turntable that supports the image inputting means 1, and changes the image pickup direction of the image inputting means 1 with two axes of the pan direction and the tilt direction. The tilt angle input means 3 is a potentiometer or the like attached to the tilt rotation shaft of the rotation means 2, and outputs a voltage signal proportional to the tilt angle θy of the image input means 1. The tracking target designating unit 4 is a console operated by an operator at the time of initial setting. The image monitor device 5 is composed of a CRT or the like, and displays an image input by the image input means 1 and an image processed by the image processing unit 6.

The image processing unit 6 includes an image storage unit 6a, a tracking gate position storage unit 6b, a target storage unit 6c, a search unit 6d, a correlation calculation unit 6e, a target position detection unit 6f, and a rotation angle. The calculation means 6g, the search area size control means 6j, and the search step interval control means 6k are provided.

The image storage means 6a is composed of an image memory or the like, and stores the image data input from the image input means 1. The tracking gate position storage means 6b stores the current position of the tracking gate. The target storage means 6c stores image data to be used as a template. The search means 6d sequentially moves the tracking gate to a position where correlation calculation is performed within a predetermined search area and at predetermined search step intervals, and obtains image data required at each moving position from the image storage means 6a. Output to 6e.

The correlation calculation means 6e performs a correlation calculation based on the image data output from the image storage means 6a by the search means 6d and the image data stored in the target storage means 6c. The target position detecting means 6f obtains the position having the highest correlation based on the respective results calculated by the correlation calculating means 6e. The rotation angle calculation means 6g calculates a rotation angle between the pan direction and the tilt direction for giving an instruction to the rotation means 2, which is necessary for adjusting the tracking target position indicated by the coordinates on the image to the image center. To do. The search area size control means 6j outputs a signal proportional to the tilt angle obtained from the tilt angle input means 3,
Convert to search area size.

The search step interval control means 6k is used for the moving dot δx per step in the horizontal direction of the tracking gate G and the vertical step of the tracking gate G in FIG. 7 used in the description of the first embodiment. Hit moving dot δ
y and y are set as follows according to the search area size set by the search area size control means 6j. That is, the search step interval control means 6k sets the values of the moving dot δx and the moving dot δy to be large when the search area size is large, and sets the moving dot δx and the moving dot when the search area size is small. The value of δy is set to a small value.

The above-mentioned automatic tracking device operates as follows. That is, the operator designates a tracking target as an initial setting before starting the automatic tracking. The image captured by the image input means 1 can be confirmed through the display screen of the image monitor device 5. The display screen displays a rectangular tracking gate indicating an area of a template for storing a tracking target. Therefore, the operator uses the tracking target designating unit 4 to operate the rotating unit 2 so that the tracking target enters the tracking gate, and starts automatic tracking when the tracking target enters the tracking gate.

When the tracking target designation operation is performed, the automatic tracking device stores the position of the tracking gate in the tracking gate position storage means 6b. Further, the automatic tracking device reads the image data in the tracking gate at the position set in the tracking gate position storage means 6b from the image storage means 6a and stores it in the target storage means 6c.

When the initial setting is completed as described above,
The automatic tracking device starts tracking. First, the image input means 1
Captures a predetermined image capturing area at regular time intervals and outputs the captured image as image data to the image storage means 6a. The image data output from the image input means 1 is the image storage means 6a.
Are sequentially stored in. The search means 6d reads the image data of the image storage means 6a corresponding to the tracking gate and outputs it to the correlation calculation means 6e. When this reading is completed, the search means 6d moves the tracking gate within the search area having the size output by the search area size control means 6j by the search step interval determined by the search step interval control means 6k.

At this time, the search area size control means 6j
Is based on the tilt angle of the image input means 1 in the imaging direction,
If the tilt angle is large, the search area searched by the tracking gate is enlarged, and if the tilt angle is small, the search area searched by the tracking gate is narrowed. This is because the larger the distance between the tracking target and the image input means 1, the shorter the moving distance of the moving tracking target on the image. That is, as shown in FIG. 9, for example, the maximum horizontal movement distance of the actual tracking target P per cycle time for performing image processing is Ux _max , the distance from the image input means 1 to the tracking target is L, and the image input is performed. The focal length of the image pickup system of the means 1 is f, and the maximum moving distance of the tracking target image on the image pickup surface is u.
When x _max , the equation (8) is established.

Ux _max = f × Ux _max / L ……………………………… (8) In other words, if Ux _max and f are known, the tracking target on the imaging surface is The maximum moving distance ux _max of the image is inversely proportional to the distance L from the image input means 1 to the tracking target. Therefore, if the distance L from the image input means 1 to the tracking target is known, the horizontal range in which the image of the tracking target on the image is searched can be limited. The vertical direction can be similarly limited. By this limitation, unnecessary search for a position where an image of a tracking target is unlikely to exist can be omitted, the processing speed for obtaining the position of the tracking target can be improved, and the influence of noise other than the search region can be eliminated.

However, although the horizontal distance L from the image input means 1 to the tracking target is not directly known, the installation height Hc of the image input means 1 is measured in advance and the tracking target height Hp is measured in advance. By setting in advance, the calculation can be easily performed by the expression (9) based on the tilt angle θy in the image pickup direction of the image input unit 1 from the tilt angle input unit 3.

L = [Hc− (Hp / 2)] / sin θy (9) Therefore, the tilt angle θy and the maximum moving distance ux _max of the tracking target image on the imaging surface are Can be related. That is, the search area size control means 6i outputs an appropriate search area that is wider as the distance L from the image input means 1 to the tracking target becomes smaller according to the tilt angle θy.

On the other hand, if the search is performed with the same search step intervals δx and δy, the larger the search area size set by the search area size control means 6i, the larger the number of steps for searching the entire search area. The processing time for calculating and determining the position of the tracking target becomes long. Therefore, the search step interval control means 6k increases the search step intervals δx and δy as the search area size set by the search area size control means 6i increases.
Instruct d to make the calculation time of the tracking target position constant.

The correlation calculation means 6e searches the entire search region at the search step intervals δx and δy set by the search step interval control means 6k as described above, and explained in the first embodiment (5). The sum σ (i, j), which is a measure of inconsistency, is calculated for each search step based on the formula. The target position detecting means 6f gives the minimum value in each sum σ (i, j) in the same manner as described in the first embodiment (i,
Based on j), the position most similar to the image in the tracking storage means 6c (template) is calculated, and the image storage means 6a is calculated.
The position of the tracking target on the image stored in is acquired.

After that, in the same way as described in the first embodiment, the rotation angle calculating means 6g is operated by the target position detecting means 6f.
The pan-direction rotation angle θx required to align the position (i, j) with the highest correlation obtained in (3) with the center of the image is set in (6) above.
Based on the equation, the tilt direction rotation angle θy is calculated as described in (7) above.
Each is calculated based on the formula and output to the rotating means 2. At this time, the target storage unit 6c newly updates and stores the image data input by the image input unit 1 to the image storage unit 6a.

The turning means 2 uses the turning angle calculating means 6g to
When a signal including the pan direction rotation angle θx and the tilt direction rotation angle θy is received, the image input means 1 is based on the signals.
The image pickup direction of is tracked and the tracking target P is rotated in the image pickup direction in which it is captured in the center of the image. Then, thereafter, the automatic tracking device realizes automatic tracking by repeating the above-described processing based on the image data newly stored in the target storage means 6c (template).

Therefore, in the automatic tracking device of the second embodiment described above, the size of the search area is changed according to the tilt angle θy of the image input means 1, so that it is unnecessary in the image. There is no waste of searching for a tracking target, and tracking performance is improved.
Further, since the search step interval is widened according to the size of the search area, even if the tracking target is located near the image input means 1 and the search area is widened, it is not necessary to reduce the tracking speed. .

[0062]

According to the invention described in claim 1, since the tracking gate size is changed according to the distance between the image input means and the tracking target, the background image of the background image is displayed regardless of the distance between the image input means and the tracking target. It is possible to provide an automatic tracking device having excellent tracking performance that can reduce the influence.

According to the second aspect of the present invention, the size of the search area is changed according to the distance between the image input means and the tracking target. Therefore, the search area is narrowed as much as possible to reduce waste of search and improve tracking accuracy. With this, it is possible to provide an automatic tracking device having excellent tracking performance capable of improving the tracking speed and the tracking speed.

According to the invention described in claim 1 , in addition to the above effects, there is an effect that it is possible to provide an excellent automatic tracking device capable of further improving the tracking speed.

According to the second aspect of the invention, in addition to the above effects, there is an effect that it is possible to provide an excellent automatic tracking device capable of further improving the tracking speed and the tracking accuracy.

[Brief description of drawings]

FIG. 1 is a block diagram showing an automatic tracking device according to a first embodiment of the present invention.

FIG. 2 is an image diagram illustrating a method of specifying a tracking target of the automatic tracking device.

FIG. 3 is an image diagram illustrating an operation of a tracking gate size control unit of the automatic tracking device.

FIG. 4 is an image diagram illustrating a relationship between a tracking target of the automatic tracking device and an image input unit.

FIG. 5 is an image diagram illustrating an operation of a tracking gate size control unit of the automatic tracking device.

FIG. 6 is an image diagram illustrating the operation of the thinning control means of the automatic tracking device.

FIG. 7 is an image diagram illustrating an operation of a search unit of the automatic tracking device.

FIG. 8 is a block diagram showing an automatic tracking device according to a second embodiment of the present invention.

FIG. 9 is an explanatory diagram showing an operation of a search area size control unit of the automatic tracking device.

FIG. 10 is an image diagram showing a relationship between a tracking gate and a tracking target of the automatic tracking device.

[Explanation of symbols]

1 Image input means 2 rotation means 4 Tracking target designation means 6f Target position detection means 6h Tracking gate size control means 6i thinning control means 6j Search area size control means 6k search step interval control means G tracking gate P Tracking target R search area δx search step interval δy Search step interval

─────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A 61-10372 (JP, A) JP-A 7-30801 (JP, A) JP-A 5-306424 (JP, A) JP-A 4- 77805 (JP, A) JP 61-12177 (JP, A) JP 61-123367 (JP, A) JP 5-103249 (JP, A) JP 56-27585 (JP, A) (58) Fields surveyed (Int.Cl. ⁷ , DB name) G01S 3/76-3/789 G01S ⁷ /48-7/50 H04N 5/232 H04N 7/18

Claims (2)

(57) [Claims] 1. A position of a tracking target by moving an image input means, a tracking target designating means for designating a tracking target from the image, and a tracking gate within a designated search area in the image at a predetermined search step interval. Target position detecting means for calculating
In an automatic tracking device including a rotating unit that rotates the image input unit according to the position of the tracking target calculated by the target position detecting unit, the tracking gate size is changed according to the rotating position in the tilt direction. Provided with gate size control means
And the tracking target position according to the turning position in the tilt direction.
Change the thinning rate of the image data used to calculate the
An automatic tracking device characterized by the addition of a control means . 2. A position of a tracking target by moving an image inputting unit, a tracking target designating unit for designating a tracking target from the image, and a tracking gate within a designated search region in the image at a predetermined search step interval. Target position detecting means for calculating
In an automatic tracking device including a rotating unit that rotates the image input unit according to the position of the tracking target calculated by the target position detecting unit, the size of the search area is changed according to the rotating position in the tilt direction. Ru provided the search area size control means
At the same time, it is specified in the image according to the rotation position in the tilt direction.
Change the search step interval to search within the search area
An automatic tracking device characterized in that a search step interval control means is added .