RU2379706C2 - Method to increase resolution of radio-and ir-images - Google Patents

Abstract

FIELD: radiolocation.

SUBSTANCE: invention relates to passive radiolocation, particularly, to radio-IR-radars (RIRR) intended for surface and air observation and built around to mobile or stationary RIRR with scanning center-aligned antennas of radiometric channels with different directional characteristics (DC). Proposed invention aims at increasing resolution of radio-IR-images in angle coordinates and expanding the range of sharp vision to RIRR field of vision. Proposed method consists in that, during surface or air observation with scanning RIRR comprising several center-aligned antennas with directional characteristics described by functions with separate variables, block matrix is formed to measure the main and mixed channels. Then, measurement matrix is multiplied, on the right and left sides, by weight factor matrix, matrix factors being calculated preliminary in compliance with a certain procedure.

EFFECT: matrix of amplitude estimates representing recovered object image in RIRR vision field with resolution in angle coordinates increased several times.

Description

The invention relates to passive radar, and in particular to radar stations (RTLS) for monitoring the surface and air conditions on the basis of mobile or stationary carriers of RTLS with scanning center-aligned antennas of radiometric channels with different characteristics of antenna patterns (NAM).

When observing RTLS for ground and air objects, line-by-line scanning by the RTLS antenna of the viewing area is carried out with a continuous displacement of the antenna in azimuth and elevation [1, pp. 71-73, 2, 3]. In this case, a two-dimensional radio thermal image (RTI) of the objects and the underlying surface is formed. The clarity of the rubber goods and the accuracy of determining the angular coordinates of objects on the rubber goods with this method of image formation is limited by the width of the bottom. Therefore, the image is blurry in angular coordinates. There is a problem of increasing the resolution of solid-state rubber products in the corners due to the algorithmic processing of the observed blurry images. Known methods for increasing the resolution in angular coordinates for scanning radars [4, 5], which can also be used to form RTI based on scanning RTLS. However, these methods have one common drawback - the limited accuracy of image restoration due to the small number of measuring channels and the loss of image information along the perimeter of the viewing area.

The closest in technical essence is a way to increase the resolution in angular coordinates [4], which consists in restoring the image matrix in range slices and reduces to the following operations.

1. The radar antenna is successively shifted in azimuth and elevation, respectively, by the value of the (2n + 1) th and (2m + 1) th part of the beam width line by line in the field of view. The beam width is 2n + 1 azimuth sampling elements and 2m + 1 azimuth sampling elements.

,

.2. The amplitudes of the reflected signals obtained at each i, jth position of the beam at the output of one total radar channel y (i, j) are formed into the measurement matrix y (i, j),

,

.

,

, искомой матрицы изображения X строится

плавающее окно размера (2m₁+1)×(2n₁+1), m₁<m, n₁<n. Амплитуды y(i+i₁, j+j₁) матрицы измерений Y, взятые в этом окне при

,

3. Regarding each i, j-th element x (i, j),

,

, the desired image matrix X is constructed

a floating window of size (2m ₁ +1) × (2n ₁ +1), m ₁ <m, n ₁ <n. The amplitudes y (i + i ₁ , j + j ₁ ) of the measurement matrix Y taken in this window for

,

summarize by i ₁ and j ₁ with weight coefficients h (i ₁ , j ₁ ), the calculation of which is carried out in advance according to a certain technique. For the BOTTOM of the total channel, which is described by a function with separated variables by angular coordinates, the weight coefficients are divided by variables: h (i, j) = h ₁ (i) · h ₂ (j), and the double sum is calculated as the repeated sum:

in the form of a double repetition of one-dimensional operations: first by rows and then by columns. Operations (1) are performed in the process of forming rows of the measurement matrix Y, which increases the speed of the image restoration algorithm X.

располагают в составе матрицы

, которая представляет собой восстановленное в данном сечении дальности изображение в i, j-x синтезированных элементах разрешения по азимуту и углу места, размеры которых в несколько раз меньше ширины ДНА.4. The received estimates

positioned as part of the matrix

, which is the image restored in this section of the range in i, jx synthesized resolution elements in azimuth and elevation, the dimensions of which are several times smaller than the width of the bottom.

5. For RTLS, the range cross section will be understood as the entire range of range in the field of view. Therefore, in contrast to the radar [4], the rubber goods will be obtained two-dimensional in the coordinates of the elevation - azimuth.

However, this method has the following disadvantages.

1. The method is limited to considering only one measurement channel and does not provide for the possibility of increasing the number of channels, which would lead to an increase in the accuracy of the estimation of amplitudes and, as a result, an increase in the resolution of images.

с элементами

теряется информация об изображении по периметру зоны обзора, и матрица

с элементами

оказывается полностью не заполненной. Это вызвано тем, что в плавающем окне вычисляется только одна оценка, соответствующая центру окна.2. In the resulting matrix of estimates

with elements

information about the image around the perimeter of the viewing area is lost, and the matrix

with elements

It turns out to be completely unfilled. This is because in the floating window, only one estimate is calculated, corresponding to the center of the window.

The technical result is aimed at increasing the resolution of the RTI in angular coordinates and expanding the area of clear vision to the field of view of the RTLS.

,

, которую далее обрабатывают, отличающийся тем, что в антенной системе число каналов k увеличивают до К, где k=1, 2, …, K, дополнительно формируют матрицы измерений амплитуд основных каналов Y₂₂, …, Y_KK, параллельно формируют матрицы измерений амплитуд смешанных каналов Y_qk, где

,

, q≠k, которые синтезируются на основе матриц Y_kk,

составляют из полученных матриц блочную матрицу Y, затем умножают блочную матрицу Y слева и справа на блочные матрицы весовых коэффициентов Н_A и Н_B, вычисляемые заранее, и формируют матрицу оценок амплитуд всей I×J-зоны обзора

представляющую собой матрицу радиотеплового изображения поверхности или воздушной обстановки в виде совокупности амплитуд сигналов излучения в элементах дискретизации угломерного пространства с возможностью их отображения на экране индикатора РТЛС.The technical result of the proposed technical solution is achieved in that a way to increase the resolution of thermal thermal images is that when observing a surface or airborne environment using a scanning RTLS with one antenna, whose DN is described by a function with separated variables, the antenna is successively shifted in azimuth and angle places, respectively, by the value of the (2n + 1) th and (2m + 1) th parts of the beam width line by line in the field of view, the amplitudes of the received signals are measured for each i, jth ant data and form from these amplitudes a measurement matrix Y ₁₁ with elements y (i, j),

,

which is further processed, characterized in that in the antenna system the number of channels k is increased to K, where k = 1, 2, ..., K, additionally matrixes of amplitudes are measured for the amplitudes of the main channels Y ₂₂ , ..., Y _KK , amplitude matrices are formed in parallel mixed channels Y _qk , where

,

, q ≠ k, which are synthesized based on the matrices Y _kk ,

make up the block matrix Y from the resulting matrices, then multiply the block matrix Y on the left and right by the block matrices of weight coefficients H _A and H _B calculated in advance, and form a matrix of estimates of the amplitudes of the entire I × J viewing area

which is a matrix of a radio thermal image of a surface or air environment in the form of a set of amplitudes of radiation signals in the discretization elements of the goniometric space with the possibility of their display on the screen of the RTLS indicator.

The method is as follows.

1. The antenna system, which in the general case is several center-aligned antennas with different characteristics of the beam, described by functions with angularly separated variables, scans the field of view line by line, shifting in azimuth and elevation, respectively, to the (2n + 1) -th and ( 2m + 1) -th part of the width of the DN.

2. The digital processing system of the signals received by the radiometer measures the amplitudes of the signals in each kth channel (k = 1, 2, ..., K), associates these amplitudes with the corresponding i, jth antenna positions, and generates measurement matrices of the main channels Y ₁₁ , Y ₂₂ , ..., Y _KK .

,

, q≠k. Эти матрицы синтезируются на основе матриц Y_kk,

основных каналов в соответствии с методикой, изложенной в расчетной части.2. In parallel, matrixes of measurements of mixed channels Y _{qk are formed} ,

,

, q ≠ k. These matrices are synthesized based on the matrices Y _kk ,

main channels in accordance with the methodology described in the calculation part.

,

, q≠k. ДН всех каналов описываются функциями с разделенными переменными.3. The operation of the antenna system, consisting of K ² channels: K main and K ² -K mixed. Mixed i, je channels have DND characteristics corresponding to the i-th angle and azimuth of the j-th main channels. After receiving and processing the signals in the mixed channels, the measurement matrices Y _{qk are formed} ,

,

, q ≠ k. The DVs of all channels are described by functions with separated variables.

с элементами

,

,

по формуле4. From the obtained matrices, a block measurement matrix Y = (Y _qk ) is compiled and the matrix of amplitude estimates is calculated

with elements

,

,

according to the formula

where H _A and H _B are in the general case block matrices of weighting coefficients, which are calculated in advance in accordance with the methodology described in the calculation part.

представляет собой восстановленное в элементах дискретизации изображение поверхности или воздушной обстановки в зоне обзора РТЛС с повышенным в несколько раз разрешением по углам, которое выводится на экран индикатора.5. The matrix

represents the image of the surface or air situation restored in the discretization elements in the RTLS viewing area with a several-fold increased angular resolution, which is displayed on the indicator screen.

Calculation of weighting matrices is as follows.

The discrete model of measurements in the q, kth channel with the i, jth position of the antenna at the output of the digital processing system of the receiving signal is described by the following total dependence:

where α _qk (i ₁ , j ₁ ) are the normalized coefficients of the channel q, k-th channel, characterizing the intensity of the signal arrival from the i ₁ , j _1st angular direction relative to (i, j); x (i, j) are the desired parameters characterizing the amplitudes of the radiation signals in i, jx bins; p _qk (i, j) is the centered Gaussian interference of the qth, kth channel: p _qk (i, j) ∈N (0, σ ² _P ), including model formation errors (3) and hardware noise.

For circular and fan-shaped antenna patterns of RTLS antennas, the coefficients of the antenna pattern can be represented as functions with separated variables:

for example: α _qk (i, j) = exp (-λ _q · i ² ) · exp (-µ _k · j ² ),

where λ _q , μ _k are the approximation parameters of the DN along the vertical and horizontal, respectively (for a circular DND, _{q q} = μ _q ).

Taking into account property (4), model (3) takes the form

,

в следующей матричной форме:which allows us to represent a set of i, jx measurements,

,

in the following matrix form:

- (J+2n)×J-матрица ленточного типа коэффициентов ДН, описывающая смазывание изображения по горизонтали, причем для круговой ДН матрицы А и В совпадают: А=В; P=(p(i, j)) - I×J-матрица помех; Т - символ транспонирования.where Y _qk = (Y _qk (i, j) is the I × J-matrix of measurements of the qth, kth channel; A = (α _si ) -I × (I + 2m) is a tape-type matrix of day-to-day coefficients describing image blurring vertically; X = (x (i, j)) - (I + m) × (J + m) -matrix of the desired parameters of the radiation field;

- (J + 2n) × J-matrix of the tape type of the coefficients of the pattern, which describes the horizontal blur of the image, moreover, for a circular pattern, the matrices A and B are the same: A = B; P = (p (i, j)) - I × J interference matrix; T is the symbol for transposition.

Placing the matrices of formula (6) in the corresponding block matrices Y, A, B, P, we obtain the following matrix-block model of measurements:

where the matrix X is multiplied by block matrices according to the Kronecker product rule [6].

In the particular case for the two main channels of the RTLS (K = 2) we have:

искомой матрицы Х можно подчинить следующему условию:Search for the best grade

the desired matrix X can be subordinated to the following condition:

измерений Y от оценок измерений

, восстановленных на основе

с точностью до ошибок измерения Р и ошибок оценивания

where the matrix function F characterizes the deviation

Y measurements from measurement estimates

recovered based

accurate to measurement errors P and estimation errors

The trace Tr [F] of matrix (9) is the sum of the squares of the deviation of the measurements of all channels from their restored values.

:The solution to problem (8) is the matrix of estimates

:

Calculation of (10) is equivalent to the two-stage procedure of the least squares method (least squares) - first, blur along the rows of the image matrix is eliminated, and then blur along the columns is eliminated:

Implementation (11) leads to the following algorithm:

where when the matrices are inverted, the regularization parameter δ is used [8, p. 55], and the matrices of the weight coefficients Н _A and H _B are calculated in advance and in the case of circular DNDs coincide: Н _A = Н _B.

:Finally, we can write the following formula for finding the matrix of estimates

:

which is equivalent to (11).

For RTLS with two main channels of expression (12) are disclosed:

и аналогично для В^T B.Where

and similarly for B ^T B.

The introduction of the regularization parameter δ during matrix inversion is carried out similarly to (12). Each stage of procedure (14) is performed in the same way, and for circular DNDs: A = B and H _A = H _B , which is convenient from a computational point of view.

Finding estimates of the measurement matrices of mixed channels can be carried out on the basis of measurements of the first channel (with the narrowest DN):

.In the case of circular DNs: A = B and

.

и

вводится параметр регуляризации δ, необходимый для обращения матриц:For ratings

and

the regularization parameter δ necessary for matrix inversion is introduced:

,

.

,

.

,

вместо реальных измерений Y₂₁, Y₁₂ снижается, поэтому целесообразней вместо синтезированных измерений

,

использовать реальные измерения Y₂₁, Y₁₂ смешанных каналов.RTI recovery accuracy (X grades) using grades

,

instead of real measurements, Y ₂₁ , Y ₁₂ decreases, therefore it is more expedient instead of synthesized measurements

,

use real measurements of Y ₂₁ , Y ₁₂ mixed channels.

Simulation results. Figure 1-3 show: figure 1 is a simulated image of the surface; figure 2 - image compressed in rows and columns by 7 times - simulation of scanning 7 × 7-day viewing area with an offset across the width of the day; figure 3 is an image reconstructed by a two-stage algorithm (14) during element-by-element scanning of the bottom of the field of view.

The proposed method allows several times to increase the resolution of RTLS in azimuth and elevation while maintaining the field of view of RTLS in azimuth and elevation and to form a matrix of the thermal image of the surface or air situation in the form of a set of estimates of the amplitudes of the radiation signals in the discretization elements of the goniometric space with their display on indicator screen. This makes it possible to monitor ground and air objects in the absence of optical visibility.

Literature

1. Nikolaev A.G., Pertsov S.V. Radiolocation. M .: Military. Publishing House., 1970.132 s.

2. Pirogov Yu.A. Passive radio vision in the millimeter range // Radio engineering. 2003. No2. S.4-11.

3. Pirogov Yu.A., Timanovsky A.L. Superresolution in systems of passive radio-vision of millimeter range // Radio engineering. 2006. No3.

4. Patent RU 2284548 C1. The method of monitoring the surface and air conditions on the basis of the airborne radar / V.K. Klochko. IPC: G01S 13/02. Priority 06.23.2005. Published: 09/27/2006. Bull. Number 27.

5. Patent RU 2292060 C1. A method for observing airborne objects and a surface based on an onboard radar / V.K. Klochko. IPC: G01S 13/02. Priority 06/28/2005. Published: 01/20/2007. Bull. No. 2.

6. Bellman R. Introduction to the theory of matrices / Per. from English M .: Science. 1976.

7. Monzingo R.A., Miller T.U. Adaptive Antenna Arrays: Introduction to Theory / Transl. from English M .: Radio and communications, 1986. 448 p.

8. Vasilenko G.I., Taratin A.M. Image recovery. M .: Radio and communication, 1986. 304 p.

Claims (1)

A way to increase the resolution of thermal thermal images, which consists in the fact that when observing a surface or airborne environment using a scanning radio thermal location station (RTLS) with one antenna, the radiation pattern (BH) of which is described by a function with separated variables, the antenna is successively shifted in azimuth and angle places, respectively, by the value of the (2n + 1) th and (2m + 1) th parts of the beam width line by line in the field of view, the amplitudes of the received signals are measured at each i, jth antenna position and form the amplitudes of these measurements Y matrix ₁₁ with elements y (i, j),

which is further processed, characterized in that in this antenna system the number of channels k is increased to K, where k = 1, 2, ..., K, additionally matrixes for measuring the amplitudes of the main channels Y ₂₂ , ..., Y _kk are formed, and matrixes for measuring the amplitudes mixed channels Y _qk , where

q ≠ k, which are synthesized based on the matrices Y _kk ,

make up the block matrix Y from the obtained matrices, then multiply the block matrix Y left and right by the block matrices of weight coefficients H _A and H _B calculated in advance, and form a matrix of estimates of the amplitudes of the entire I × J viewing area

which is a matrix of a radio thermal image of a surface or air environment in the form of a set of amplitudes of radiation signals in the discretization elements of the goniometric space with the possibility of their display on the screen of the RTLS indicator.

Images