A method of tracking a rotating radar target

Abstract

The application discloses a kind of rotating radar target tracking methods, belong to target tracking technical field, solve the problem that the tracking method in prior art is prone to lose tracking target and lead to unstable tracking.The target tracking method includes: step 1, obtains the imaging result of two antennas of double-channel SAR system, interpolates and registers two images therein, compensates the phase error caused by the deviation of receiving aperture azimuth position;Step 2, based on the obtained slow-time compensation imaging result, obtain micro-motion target energy, determine target position based on slow-time micro-motion target energy;Step 3, track the detected target.The application can be used for rotating platform target tracking guidance, realizes the stable tracking of target.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of target tracking, and in particular to a rotating radar target tracking method. BACKGROUND

[0002] In recent years, the demand for automatic tracking of moving targets in complex backgrounds is increasingly urgent. At the same time, as the carrying platform of the tracking device, the rotation of the target in the image often leads to the loss of target tracking, and the existing target tracking method cannot meet the current application requirements. SUMMARY

[0003] In view of the above analysis, the present application aims to provide a rotating radar target tracking method to solve the technical problems of the existing tracking method that is not very accurate and easy to lose the target.

[0004] The purpose of the present application is mainly realized through the following technical solutions:

[0005] The present application provides a rotating radar target tracking method, comprising the following steps:

[0006] Step 1, obtaining the imaging results of the two antennas of the dual-channel SAR system, interpolating and registering two images among them, and compensating for the phase error caused by the deviation of the receiving aperture position in the azimuth direction;

[0007] Step 2, obtaining the micro-motion target energy, and processing the micro-motion target energy to obtain the target image;

[0008] Step 3, tracking the detected target.

[0009] Further, step 3 comprises the following sub-steps:

[0010] Step 31, obtaining an initial frame image for target tracking; the initial frame image and the subsequent frame image have a rotation angle;

[0011] Step 32, obtaining a subsequent frame image, rotating the image according to the rotation angle existing with the initial frame image to obtain a rotated image eliminating the rotation angle;

[0012] Step 33, taking the target point coordinates of the previous frame image as the center, using a correlation filtering model to detect the target in the rotated image to obtain the target point coordinates in the rotated image;

[0013] Step 34, converting the target point coordinates in the rotated image to the target point coordinates in the pre-rotated image, and calculating the line-of-sight angular velocity of the rotating platform as the tracking output signal according to the target point coordinates;

[0014] Step 35, taking the target point coordinate of the rotated image as the center to intercept the search area to update the correlation filter model, and according to the method of steps 32-34, tracking the target in the next frame image;

[0015] Repeat steps 32-35 until all subsequent frame images are tracked.

[0016] Further, in step 32, according to the rotation angle existing with the initial frame image, the image is rotated to obtain the pixel point coordinate (x', y') of the rotated image eliminating the rotation angle as: Where (x, y) is the pixel point coordinate of the pre-rotated image; (x r ,y r ) is the coordinate of any point on the pre-rotated image as the rotation axis; and ξ is the image rotation angle.

[0017] 4. The target tracking method of rotating radar according to claim 3, characterized in that step 33 comprises:

[0018] 1) interpolating the rotated image to obtain a gray scale image including the gray scale values of each pixel point in the image;

[0019] 2) taking the target point coordinate of the previous frame image as the center, intercepting a set size image in the gray scale image after normalization, and extracting the Hog feature to obtain the feature matrix Z;

[0020] 3) using the correlation filter model to detect the target in the intercepted set size image to obtain the target point coordinate in the rotated image.

[0021] Further, step 34 comprises:

[0022] 1) converting the target point coordinate in the rotated image to the pre-rotated image to obtain the target point coordinate (X T , Y T ) of the target point in the pre-rotated image;

[0023] 2) calculating the line-of-sight angular velocity of the rotating platform as the tracking output signal according to the formula

[0024] Where V x and V y are the X direction and Y direction line-of-sight angular velocities, X c and Y c are the center point coordinates of the X direction and Y direction of the pre-rotated image, R x and R y are the X direction and Y direction line-of-sight angular velocity equivalent values.

[0025] Further, in step 34, the target point coordinate (X' T ​, Y' T ) into the pre-rotation image, to obtain the target point coordinate (X T , Y T ) of the target point in the pre-rotation image.

[0026] Wherein, (x, y) is the coordinate of the point in the pre-rotation image; (x', y') is the pixel coordinate of the post-rotation image; (x r , y r ) is the coordinate of any point in the pre-rotation image as the rotation axis; ξ is the image rotation angle.

[0027] Further, in step 1, the two complex images are interpolated and registered by using a compensation function C 12 (t a ), which compensates the phase error caused by the deviation of the receiving aperture azimuth position, and the compensation function C 12 (t a ) is:

[0028]

[0029] Wherein, j is the imaginary unit, k is the signal frequency; R0 is the distance from the radar to the target; t a is the time of flight; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis.

[0030] Further, in step 2, the phases of the two images are extracted, the interference phase is detected by setting a certain threshold phase, the stationary target is cancelled, and then the cancellation result is taken modulo to obtain the micro-motion target energy.

[0031] Further, in step 2, the stationary target is cancelled, and the cancellation result is:

[0032]

[0033] Wherein, j is the imaginary unit; k is the signal frequency; R0 is the distance from the radar to the target; t a is the slow time; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis.

[0034] Further, in step 2, the square of the modulo value of the cancellation result is taken to obtain the micro-motion target energy:

[0035]

[0036] wherein, A3(t r ,t a ) is an additional modulation term; λ is the wavelength of the radar transmitted signal; k is the signal frequency modulation; R0 is the distance from the radar to the target; r is the target rotation radius; r << R o ; ω is the rotation speed; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis; B is the signal bandwidth; c is the speed of light; t r is the fast time; t a is the slow time, J m (2kr) is the first type of m-order Bessel function, and m is the order; B d is the Doppler bandwidth of the echo signal.

[0037] Compared with the prior art, the present application can achieve at least one of the following beneficial effects:

[0038] (1) The method has a wide range of application and can be universally applied to photoelectric tracking guidance in high-speed rotating platforms.

[0039] (2) The method has low complexity and good real-time performance, and can meet the real-time calculation requirements on a commonly used embedded hardware platform.

[0040] (3) The present application can use the change of the Doppler frequency of the echo of different antennas to offset the stationary target and extract the moving target, and use the relationship between the phase difference of the two SAR images and the system parameters of the interferometric SAR and the target motion parameters to obtain the target motion parameter information.

[0041] The above technical solutions can be combined with each other in the present application to achieve more preferred combination solutions. Other features and advantages of the present application will be described in the subsequent specification, and some advantages will become apparent from the specification, or will be understood by implementing the present application. The purposes and other advantages of the present application can be achieved and obtained through the contents specifically indicated in the specification, examples and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0042] The accompanying drawings are included to provide a further understanding of the present application and are incorporated herein and constitute a part of the specification. The drawings illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application. In the drawings:

[0043] Figure 1 is a flowchart of the inversion positioning method of the present application;

[0044] Figure 2 is a schematic diagram of the dual-channel SAR system of the present application;

[0045] Figure 3 is a schematic diagram of the ship and micro-motion target imaging result of the present application;

[0046] Figure 4 Fig. 1 is a schematic diagram of micro-motion target interference detection results of the present application;

[0047] Figure 5 Fig. 2 is a schematic diagram of micro-motion target repositioning results of the present application;

[0048] Figure 6 Fig. 3 is a schematic diagram of target tracking flow of the present application. DETAILED DESCRIPTION

[0049] The preferred embodiments of the present application will be described in detail below with reference to the accompanying drawings, which form a part of this application. The accompanying drawings and the detailed description below are used to explain the principles of the present application and are not intended to limit the scope of the present application.

[0050] The present application provides a target tracking method based on multi-channel detection, comprising the following steps:

[0051] Step 1, obtaining imaging results of two antennas of a dual-channel SAR system, interpolating and registering two images therein, and compensating phase errors caused by azimuth position deviation of a receiving aperture

[0052] Step 2, obtaining micro-motion target energy based on the obtained imaging results after slow time compensation, and determining target position based on the slow time micro-motion target energy; Step 3, tracking the detected target.

[0053] It should be noted that in the above step 1, the process of obtaining imaging results comprises the following sub-steps:

[0054] Step 11, defining that the signal transmitted by the No. 2 antenna of the dual-channel SAR system is received by the No. 1 antenna and the No. 2 antenna at the same time, establishing a radar coordinate system (X, Y, Z) with the No. 2 antenna as the origin, and the origin is always fixed on the phase center of the No. 2 antenna;

[0055] Step 12, sub-aperture off-loading frequency and fast time matched filtering processing;

[0056] Step 13, compensating additional Doppler centroid deviation caused by azimuth deviation of the dual-channel antenna;

[0057] Step 14, azimuth processing by using azimuth reference function convolution to obtain imaging results;

[0058] As Figure 1As shown, the present application firstly performs SAR imaging on two echo signals respectively; then interpolates and registers two images; subsequently multiplies two image data conjugately to extract phase; detects the interference phase by setting a certain threshold phase, and determines that the target exists if the interference phase exceeds the threshold phase; finally obtains target motion parameter information by using the relationship between the phase difference of two SAR images and the interference SAR system parameters and target motion parameters.

[0059] Compared with the prior art, the present application uses the change of echo Doppler frequency of different antennas to offset stationary targets and extract moving targets; uses the relationship between the phase difference of two SAR images and the interference SAR system parameters and target motion parameters to obtain target motion parameter information. The present application combines the high-resolution imaging of ground targets by SAR and the detection function of micro-motion targets, processes the influence of the micro-motion effect of targets, and labels the influence on clear SAR images, can detect, locate and track high-value targets, can provide favorable information for on-site situation assessment, command and control, and has very important significance for reconnaissance and on-site perception.

[0060] In step 11 above, as shown, Figure 2 R0 is the distance from the radar to the target, the target rotation radius is r, r << R o , the rotation speed is ω, and the initial phase is t a is the flight time, D a is the distance between the phase centers of two antennas of a dual-channel SAR system, antenna No. 2 transmits, and echo signals are received by antenna No. 1 and antenna No. 2 at the same time, a radar coordinate system (X, Y, Z) is established with antenna No. 2 as the origin, the origin is always fixed on the phase center of antenna No. 2, and the SAR platform flies along the positive direction of the X axis at a speed V a ; the position of the target P in the radar coordinate system is (x0, y0, -h), the flight speed of the SAR platform is (V a , 0, 0), and the phase center position of antenna No. 1 is (0, 0, D a ).

[0061] It should be noted that, Figure 2 the dashed line represents a plane parallel to the XOY plane.

[0062] In step 12 above, the two sub-apertures are de-chirped by orthogonal demodulation, and matched filter processing is performed in the fast time domain, and the obtained echo signal is:

[0063]

[0064] wherein, Doppler center frequency and Doppler rate of the uniformly rotating passive micro-motion jamming echo received by the middle aperture; λ is the wavelength of the radar transmitted signal; j is an imaginary number; k is the signal Doppler rate; R0 is the distance from the radar to the target; r is the target rotation radius; r << R o ; ω is the rotation speed; is the initial phase; t a is the time of flight; is the corresponding position history; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis; B is the signal bandwidth; T p is the signal pulse width; c is the speed of light; t r is the fast time; t a is the slow time, m is a constant, and θ(t a ) is the transformation angle.

[0065] In the above step 13, due to the azimuth deviation between the dual-channel antennas, an additional Doppler centroid deviation is caused, and therefore compensation is needed. The compensation function is as follows:

[0066]

[0067] wherein j is an imaginary number; k is the signal Doppler rate; R0 is the distance from the radar to the target; t a is the time of flight; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis.

[0068] The compensation is directly performed by multiplying the compensation function C1(t a ) with J p1 in the above step, and J p2 remains unchanged. The specific compensation process is as follows:

[0069]

[0070] That is, the echo signal of the antenna 1 after compensation is as follows:

[0071] wherein Doppler center frequency and Doppler rate of the uniformly rotating passive micro-motion jamming echo received by the middle aperture; λ is the wavelength of the radar transmitted signal; j is an imaginary number; k is the signal Doppler rate; R0 is the distance from the radar to the target; r is the target rotation radius; r << R o ; ω is the rotation speed; is the initial phase; t a is the time of flight; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; Va V is the flight speed of the SAR carrier along the positive direction of the X axis; B is the signal bandwidth; Tp is the signal pulse width; c is the speed of light; t r is the fast time; t a is the slow time, J m (2kr) is the first kind of m-order Bessel function, and m is the order; B d is the Doppler bandwidth of the echo signal, and θ(t a ) is the transform angle.

[0072] In the above step 14, the azimuth reference function is:

[0073]

[0074] Convolution processing is performed on the above azimuth phase reference function with J p1 and J p2 respectively to obtain the imaging result

[0075]

[0076] wherein A3(t r ,t a ) is an additional modulation term, which reflects the influence of the additional phase term on the image expansion, blurring, etc., and does not affect the peak position of the interference; θ(t a ) is the transform angle; λ is the wavelength of the radar transmitted signal; j is an imaginary number; k is the signal frequency modulation; R0 is the distance from the radar to the target; r is the target rotation radius; r o << R a ; ω is the rotation speed; is the initial phase; t a is the flight time; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis; B is the signal bandwidth; T p is the signal pulse width; c is the speed of light; t r is the fast time; t a is the slow time, J m (2kr) is the first kind of m-order Bessel function, and m is the order; B d is the Doppler bandwidth of the echo signal, and θ(t a ) is the transform angle.

[0077] In the above step 1, the following compensation function is used to interpolate and register the obtained two images, so as to compensate the phase error caused by the azimuth position deviation of the receiving aperture.

[0078] Specifically, after the imaging processing, since there is an azimuth position deviation of the receiving aperture, the phase error caused by this position deviation needs to be compensated before the ground clutter cancellation.

[0079] The compensation function is shown in the following formula:

[0080]

[0081] The compensation function C 12 (t a ) is multiplied by I P1 , and the result is approximately equal to I P2 ; the specific calculation process is as follows:

[0082]

[0083] I p1 (t r ,t a )*C 12 (t a )≈I p2 (t r ,t a )

[0084] It should be noted that in the above calculation process, x0=V a *t a .

[0085] In the formula, The term actually corresponds to the compensation function The compensation function C1(t a ) is because the time-domain first term corresponds to the frequency-domain shift, and in order to align the center, compensation is performed, but after imaging, a double time-domain image is obtained, so this phase before compensation corresponds to the phase error caused by the antenna aperture, which can be seen to be irrelevant to the starting position x0 of the target point; where j is the imaginary unit, k is the signal frequency; R0 is the distance from the radar to the target; t a is the time of flight; D a is the distance between the phase centers of the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis.

[0086] In the above step 2, the phases of the two images are extracted, the interference phase is detected by setting a certain threshold phase, the stationary target is canceled, that is, the two images are subtracted and canceled, and then the cancellation result is taken modulo (see the following formula) to obtain the energy of the micro-motion target;

[0087]

[0088] The square of the modulo value of the cancellation result is obtained to obtain the energy of the micro-motion target:

[0089] Where, A3(t r ,t a) is the additional modulation term; λ is the wavelength of the radar transmission signal; k is the signal modulation frequency; R0 is the distance from the radar to the target; r is the target rotation radius; r<<R o ;ω is the speed; t a is the flight time; D a is the phase center distance between the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis; B is the signal bandwidth; c is the speed of light; t r Fast time; t a For slow time, J m (2kr) is the first kind of m-order Bessel function, where m is the order; B d is the Doppler bandwidth of the echo signal, θ(t a ) is the transformation angle.

[0090] After obtaining the energy of the micro-motion target, the maximum value of the micro-motion target energy is calculated, and the position of the maximum value is the position of the micro-motion target in the SAR image.

[0091] In one embodiment, through digital simulation, a ship is selected as an imaging target, wherein the ship warning radar has micro-motion target characteristics, and the simulation parameters used are shown in Table 1.

[0092] Table 1 Simulation parameters of micro-motion target

[0093]

[0094] According to the above simulation parameters, the ship target imaging results are as follows: Figure 3 As shown in Figure 2, we can see that the ship target can be well imaged and the outline of the ship is clear. At the same time, during the accumulation time, the radar rotation causes obvious Doppler stripes, such as Figure 3 As shown in the white circle.

[0095] The method provided by the present invention is used to perform dual-channel image interference cancellation, and to detect micro-movement targets based on the detection threshold, and to extract micro-movement target results such as Figure 4 According to the relationship between the dual-channel interferometric phase and the rotation parameters, the relocation of the rotating radar target is realized, and the results are shown in Figure 5 As shown in the figure, when the SNR is better than 15dB, the corresponding target positioning accuracy can reach 1 / 15 of the beam width, and is better than 0.5° when the beam width is less than 7°.

[0096] It should be emphasized that after steps 1 and 2 of the present invention detect and locate the target, a series of imaging results at consecutive moments are obtained; the previous image at adjacent moments is defined as the initial frame image, and the image at the next moment is defined as, based on the obtained series of imaging results at consecutive moments, the present invention can also track the target, that is, it also includes step 3, the process of tracking the target.

[0097] As Figure 6 shown, the specific tracking process for tracking the detected target includes:

[0098] Step 31, obtaining an initial frame image for target tracking; the initial frame image has a rotation angle with a subsequent frame image;

[0099] Step 32, obtaining a subsequent frame image, performing image rotation according to the rotation angle existing with the initial frame image to obtain a rotated image eliminating the rotation angle;

[0100] Step 33, taking the target point coordinates of the previous frame image as the center, using a correlation filtering model to perform target detection in the rotated image to obtain the target point coordinates in the rotated image; or, directly using the target positioning coordinates obtained in step 2.

[0101] Step 34, converting the target point coordinates in the rotated image to the target point coordinates in the pre-rotated image, and calculating the line-of-sight angular velocity of the rotating platform as a tracking output signal according to the target point coordinates;

[0102] Step 35, taking the target point coordinates of the rotated image as the center to intercept a search area to update the correlation filtering model, and performing target tracking on a further subsequent frame image according to the method of steps 32-34;

[0103] Repeat steps 32-35 until all subsequent frame images are tracked.

[0104] Specifically, since the rotating platform used in the embodiment performs target tracking, there is a certain image rotation angle between each frame image due to the selection of the platform.

[0105] Specifically, the calculation of the rotation angle in the embodiment includes:

[0106] 1) According to the attitude data of the rotating platform, a transformation matrix T from a point p w (x w , y w , z w ) in the world coordinate system to a point p m (x m , y m , z m ) in the imaging coordinate system is established:

[0107]

[0108] Where φ, θ, γ are the imaging heading angle, pitch angle, and roll angle, respectively. T -1 is the transformation matrix from the imager coordinate system to the geodetic coordinate system:

[0109] 2) Inverse the transformation matrix T to get T -1 get the transformation matrix from the imager coordinate system to the geodetic coordinate system;

[0110]

[0111] 3) get the image rotation angle ξ:

[0112]

[0113] Specifically, in step 32, according to the rotation angle existing with the initial frame image, the pixel point coordinates (x', y') of the rotated image after eliminating the rotation angle are obtained by image rotation: where (x, y) is the pixel point coordinates of the image before rotation; (x r , y r ) is the coordinates of any point on the image before rotation as the rotation axis; and ξ is the image rotation angle.

[0114] Specifically, the step 33 includes:

[0115] Step 3301, interpolation is performed on the rotated image to obtain a gray scale image including the gray scale values of each pixel point in the image;

[0116] The image coordinates obtained after rotation are generally floating-point numbers, while the image pixel coordinates are generally integers. When the image is rotated, bilinear interpolation is used to obtain the gray scale value of the to-be-sampled point by linear interpolation in two directions using the gray scale values of the surrounding four neighboring points, that is, the gray scale value of the to-be-sampled point is calculated according to the distance between the to-be-sampled point and the neighboring points. Wherein (x, y) coordinates represent the position of the pixel, and f(x, y) represents the gray scale value of the pixel, and the mathematical expression is:

[0117] f(i+u,j+v)=(1-u)(1-v)f(i,j)+(1-u)vf(i,j+1)+u(l-v)f(i+1,j)+uvf(i+1,j+1);

[0118] In the above formula, u and v take values in the range of (0, 1), representing the difference between the x coordinate and y coordinate of the upper left corner point of the four neighboring points and the to-be-sampled point; i and j represent the x coordinate and y coordinate of the upper left corner point of the four neighboring points and the to-be-sampled point.

[0119] Step 3302, before the target point coordinates of the previous frame image are taken as the center, a set size of the image is intercepted in the gray scale image, and after normalization, the Hog feature is extracted to obtain a feature matrix Z;

[0120] Step 3303, a correlation filtering model is used to detect the target in the image of a set size to obtain the target point coordinates in the rotated image.

[0121] The target capture of the initial frame image can adopt target edge extraction plus target recognition, or learning training classifier, and existing target recognition technologies such as training neural network.

[0122] Specifically, in step 3302, the rotated image is I, the initial size of the target is (w, h), and the image I' of size (4w, 4h) is cut from the image I with the target point coordinates of the previous frame image as the center. The image I' is normalized to an image I" of size 64*64, and the steps of extracting the Hog feature of the image I" are as follows:

[0123] 1) Calculate the X and Y direction gradients, including gradient amplitude and angle;

[0124] x, y direction gradient value:

[0125]

[0126] Gradient amplitude and gradient angle:

[0127]

[0128] In the above formula, f(x+1, y), f(x-1, y), f(x, y+1), f(x, y-1) respectively represent the gray value of the corresponding position pixel, G x (x, y) represents the x direction gradient value at the coordinate (x, y), G y (x, y) represents the y direction gradient value at the coordinate (x, y), G(x, y) is the gradient amplitude, and θ(x, y) is the gradient angle.

[0129] 2) Divide the cell and construct the gradient direction histogram of each cell;

[0130] Specifically, a soft mapping strategy is adopted, and when constructing the gradient direction histogram, the pixel at the junction position of multiple cells is distributed to the adjacent cells in a linear interpolation manner;

[0131] More specifically, P1-P4 represent four pixel point positions, corresponding to four linear interpolation cases:

[0132] P1 is shared by four cells;

[0133] P2 is shared by the upper and lower two cells;

[0134] P3 is shared by the left and right two cells;

[0135] P4 is only mapped to one cell.

[0136] Mapping relationship of P1 point:

[0137] Suppose the coordinates of P1 are (x, y), then the mapping formula of the gradient histogram vector of P1 point to the adjacent 4 cells is:

[0138]

[0139] In the above formula, cellsize represents the number of pixels contained in each cell, and respectively represent the downward rounding of a and b, h p1 represents the gradient histogram vector of P1 point, h A , h B , h C , h D respectively represent the gradient histogram vectors of the 4 cells adjacent to P1 point; preferably, the value of cellsize is 4.

[0140] The mapping formula of P2 and P3 points is:

[0141]

[0142] In the above formula, the calculation method of a and b is the same as the previous formula, h p2 represents the gradient histogram vector of P2 point, h p3 represents the gradient histogram vector of P3 point, h E , h F respectively represent the gradient histogram vectors of the upper and lower 2 cells adjacent to P2 point, h G , h H respectively represent the gradient histogram vectors of the left and right 2 cells adjacent to P3 point.

[0143] The mapping formula of P4 point is:

[0144] h I + = h P4 ;

[0145] In the above formula, h p4 represents the gradient histogram vector of P4 point, h I represents the gradient histogram vector of the cell where P4 point is located.

[0146] 3) Normalize the gradient histogram, and connect the normalized gradient histogram vectors together to obtain the Hog feature matrix.

[0147] Specifically, in the normalization of the gradient histogram, each cell is normalized, 2*2 cells are integrated into one block, and the L2 norm of the undirected gradient histogram vector is used to calculate the normalization factor of each block.

[0148] The cell unit E (coordinates (i, j)) belongs to four blocks B1, B2, B3, and B4 in space, and the normalization factor of each block is:

[0149]

[0150] In the above formula, δ and γ take values of 1 or -1, The gradient histogram vector of the cell unit E, N δ，γ (i, j) is the block normalization factor of the cell unit E at the position (δ, γ).

[0151] Preferably, when δ and γ both take the value of 1, N δ，γ (i, j) is the normalization factor of the cell unit E in the lower right block, The gradient histogram vector of the cell unit E on the right, The gradient histogram vector of the cell unit E below, The gradient histogram vector of the cell unit E in the lower right.

[0152] The elements of each gradient histogram vector are normalized using the normalization factor, data is truncated using a set threshold, and the average of the four values obtained after normalizing each element is taken as the final result of the current element after normalization.

[0153] Specifically, the elements of each gradient histogram vector are normalized using four normalization factors, truncated using a threshold of 0.2, and the average of the four values obtained after normalizing each element is taken as the final result of the current element after normalization, as shown in the following formula:

[0154]

[0155] In the above formula, is the gradient histogram vector of the cell unit E in the kth quadrant after averaging the gradient direction division at the coordinates (i, j), represents the gradient histogram vector of the cell unit E at the coordinates (i, j) before averaging; N -1，-1 (i, j), N 1，-1 , N -1，1 , N 1,1 represent the normalization factors of the cell unit E in the upper left, upper right, lower left, and lower right blocks, respectively; T0.2 (x) represents threshold truncation of the variable x in the parentheses, and average() represents average of multiple values in the parentheses;

[0156] T 0.2 (x) represents threshold truncation of the variable x in the parentheses, and average() represents average of multiple values in the parentheses;

[0157] In addition, the undirected gradient normalized by different normalization factors is accumulated to obtain four gradient energy sums, respectively, as the texture distribution features of a cell, and the formula is as follows:

[0158]

[0159] In the above formula, nbins represents the number of quadrants divided by all gradient directions, and represents the cumulative sum of all directional gradient values; preferably, nbins is 6.

[0160] The normalized histogram vectors described above are connected together to obtain a Hog feature matrix.

[0161] Specifically, in the target detection in the image of a set size in step 3303, a correlation filter model is used to calculate a filter response matrix R; and the target point coordinates (X T ,Y′ T ) are obtained by searching for a peak value in the filter response matrix R.

[0162] Wherein, F -1 is the inverse Fourier transform; is the Fourier transform of the correlation filter model A; is the Fourier transform of the Gaussian kernel function; λ is a regularization parameter; is the Fourier transform of the cross-correlation function; the regularization parameter λ is set to 0.1.

[0163] Correlation function σ is the width parameter of the Gaussian kernel function; preferably, the σ of the Gaussian kernel function is set to 0.125. X is the template data of the correlation filter model. is the conjugate of the Fourier transform of the template data; is the Fourier transform of the feature matrix Z.

[0164] Specifically, the step 34 includes:

[0165] 1) converting the target point coordinates in the rotated image to the pre-rotation image to obtain the target point coordinates (X T ,Y T ) of the target point in the pre-rotation image;

[0166] The target point coordinates (X' T ,Y' T ) in the rotated image are converted into the pre-rotated image, and the formula for the target point coordinates (X T ,Y T ) of the target point in the pre-rotated image is as follows:

[0167] where (x, y) is the coordinates of the point in the pre-rotated image; (x', y') is the pixel coordinates of the rotated image; (x r ,y r ) is the coordinates of an arbitrary point in the pre-rotated image as the rotation axis; and ξ is the image rotation angle.

[0168] 2) The line-of-sight angular velocity of the rotating platform is calculated as the tracking output signal according to the formula

[0169] where V x and V y are the X-direction and Y-direction line-of-sight angular velocities, X c and Y c are the X-direction and Y-direction center point coordinates of the image, R x and R y are the X-direction and Y-direction line-of-sight angular velocity equivalent values.

[0170] Specifically, the updating of the correlation filtering model in step 35 includes template data updating and correlation model updating.

[0171] The template data updating includes:

[0172] Before the correlation filtering model is updated, the initial value of the template data X of the correlation filtering model is obtained by extracting the Hog feature matrix from the image of a set size in the gray-scale image of the initial frame image after the image is centered on the target point coordinates of the initial frame image and normalized.

[0173] The updating formula of the template data X of the correlation filtering model is X new = (1-α) * X old + α * X update ; where α is the updating rate; X old is the template data of the last correlation filtering; and X update is the Hog feature matrix calculated by centering on the target point coordinates of the rotated image of the latest frame image that has been tracked and then cutting and searching the region.

[0174] The correlation model updating includes:

[0175] Before the correlation filtering model is updated, the initial value of the template data X of the correlation filtering model is obtained by extracting the Hog feature matrix from the image of a set size in the gray-scale image of the initial frame image after the image is centered on the target point coordinates of the initial frame image and normalized. ​is the Fourier transform of the Gaussian kernel function; λ is a regularization parameter; K XX is the autocorrelation function of the initial value of the template data X;

[0176] The update formula of the correlation filter model is A new = (1-α) * A old + α * A update ; wherein, α is an update rate; A old is the correlation filter model of the last correlation filtering; A update is obtained by using the updated template data X,

[0177]

[0178] α is an update rate, and the value is set to 0.012.

[0179] For example, when the correlation filtering of the initial frame image and the subsequent frame image is performed, the correlation filter model is in the state before the update, at this time, the template data X of the correlation filter model is the initial value; the correlation filter model

[0180] When the subsequent frame image is tracked after the initial frame image and the subsequent frame image are tracked, the template data and the correlation model are updated. At this time, the template data of the correlation filter model is updated, so that the template data X new = (1-α) * X old + α * X update ; wherein, X old is the template data of the last correlation filtering, and X update is the Hog feature matrix calculated by taking the target point coordinates of the rotated image of the subsequent frame image of the initial frame image as the center to intercept a search region. The correlation filter model A new = (1-α) * A old + α * A update ; wherein, α is an update rate; A old is the correlation filter model of the last correlation filtering; A update is obtained by using the updated template data X new to obtain the correlation filter model,

[0181] According to the above update process, the target tracking of all subsequent frame images is performed through the updated correlation filter model.

[0182] In summary, the stable tracking in the image rotation process is realized.

[0183] The above merely describes preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A target tracking method for a rotating radar, characterized in that: The following steps are involved: Step 1: Obtain imaging results of two antennas of a dual-channel SAR system, interpolate and align two of the images, and compensate for the phase error caused by the azimuth position deviation of the receiving aperture; Step 2: Based on the obtained slow-time compensated imaging result, obtain the micro-motion target energy, and determine the target position based on the slow-time micro-motion target energy; Step 3: Track the detected target; The following sub-steps are included: Step 31: Acquire an initial frame image for target tracking; the initial frame image and subsequent frame images have a rotation angle; Step 32: Acquire a subsequent frame image, and rotate the image according to the rotation angle with respect to the initial frame image to obtain a rotated image with the rotation angle eliminated; Step 33: Using the target point coordinates of the previous frame image as the center, a correlation filter model is used to perform target detection in the rotated image to obtain the target point coordinates in the rotated image; Step 34: Convert the coordinates of the target point in the rotated image to the image before rotation to obtain the coordinates of the target point in the image before rotation; and calculate the line of sight angular velocity of the rotating platform as a tracking output signal based on the target point coordinates. Step 35: intercepting a search area with the target point coordinates of the rotated image as the center and updating the correlation filter model, and performing target tracking on the next frame of image according to the method of steps 32-34; Repeat steps 32-35 until target tracking is performed on all subsequent frame images.

2. The target tracking method of the rotating radar according to claim 1, characterized in that: In step 32, the image is rotated according to the rotation angle with respect to the initial frame image, and the pixel coordinates (x′, y′) of the rotated image after eliminating the rotation angle are obtained as follows: Among them, (x, y) is the pixel coordinate of the image before rotation; (x r ,y r ) is the coordinate of any point on the image before rotation, and ξ is the image rotation angle.

3. The target tracking method of the rotating radar according to claim 2, characterized in that: The step 33 includes: 1) Interpolating the rotated image to obtain a grayscale image including the grayscale value of each pixel in the image; 2) Taking the target point coordinates of the previous frame image as the center, cut out an image of a set size from the grayscale image, normalize it, and extract the Hog features to obtain the feature matrix Z; 3) Use the correlation filter model to perform target detection in the image of a set size to obtain the coordinates of the target point in the rotated image.

4. The target tracking method of the rotating radar according to claim 3, characterized in that: The step 34 includes: 1) Convert the coordinates of the target point in the rotated image to the image before rotation, and obtain the coordinates of the target point in the image before rotation (X T , Y T ); 2) According to the formula Calculate the line-of-sight angular velocity of the rotating platform as the tracking output signal; Among them, V x and V y is the angular velocity of the line of sight in the X and Y directions, X c and Y c is the coordinate of the center point in the X and Y directions of the image before rotation, R x and R y is the equivalent value of the line of sight angular velocity in the X and Y directions.

5. The target tracking method of the rotating radar according to claim 4, characterized in that: In step 34, the coordinates of the target point (X′) in the rotated image are T , Y′ T ) is converted to the image before rotation, and the target point coordinates (X T , Y T ) is: Among them, (x, y) is the coordinate of the point in the image before rotation; (x′, y′) is the pixel coordinate of the image after rotation; (x r ,y r ) is the coordinate of any point in the image before rotation; ξ is the image rotation angle.

6. The target tracking method of a rotating radar according to any one of claims 1 to 5, characterized in that: In step 1, the compensation function C is used 12 (t a ) interpolates and registers the two complex images to compensate for the phase error caused by the position deviation of the receiving aperture. The compensation function C 12 (t a )for: Where j is the imaginary unit, k is the signal modulation frequency; R0 is the distance from the radar to the target; t a is the flight time; D a is the phase center distance between the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X-axis.

7. The target tracking method of the rotating radar according to claim 6, characterized in that: In step 2, the phases of the two images are extracted, and the interference phase is detected by setting a certain threshold phase to cancel the stationary target. Then, the cancellation result is modulo-measured to obtain the energy of the slightly moving target.

8. The target tracking method of the rotating radar according to claim 7, characterized in that: In step 2, the stationary target is cancelled, and the result after cancellation is: Where j is the imaginary unit; k is the signal modulation frequency; R0 is the distance from the radar to the target; t a For slow time; D a is the phase center distance between the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis; t r For quick time.

9. The target tracking method of the rotating radar according to claim 8, characterized in that: In step 2, the square of the modulus value of the cancellation result is taken to obtain the micro-motion target energy: Among them, A3(t r ,t a ) is the additional modulation term; λ is the wavelength of the radar transmission signal; k is the signal modulation frequency; R0 is the distance from the radar to the target; r is the target rotation radius; r<<R o ;ω is the rotation speed; D a is the phase center distance between the two antennas of the dual-channel SAR system; V a is the flight speed of the SAR carrier along the positive direction of the X axis; B is the signal bandwidth; c is the speed of light; t r Fast time; t a For slow time, J m (2kr) is the first kind of m-order Bessel function, where m is the order; B d is the Doppler bandwidth of the echo signal.

Images