Ship moving target detection and track generation method based on GEO SAR image sequence

Abstract

The invention relates to a ship moving target detection and track generation method based on a GEO SAR image sequence, and belongs to the technical field of target detection. GEO SAR ship imaging data are acquired, and full aperture time is divided into a plurality of sub-aperture time; imaging the GEO SAR ship imaging data according to a plurality of sub-aperture times to obtain a sub-aperture image sequence; performing K distribution CFAR detection on the sub-aperture image sequence to obtain a coarse target detection result; calculating the mass center position of the sub-aperture image sequence based on the rough target detection result; acquiring a ship track and estimated motion parameters based on the mass center position of the sub-aperture image sequence; and performing secondary coherent accumulation on the GEO SAR ship imaging data based on the motion parameters to obtain a final ship moving target focusing imaging result. According to the method, the problems that a ship target cannot be coherently accumulated for a long time and the target sequence is difficult to detect in a GEO SAR system under ultra-long synthetic aperture time are solved, and the image focusing precision is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of target detection, and in particular relates to a method for detecting moving targets of ships and generating tracks based on GEO SAR image sequences. Background Art

[0002] GEO SAR is a powerful, wide-coverage, long-term beam dwelling system for observation scenes and targets, and is suitable for ship target detection and tracking. The long synthetic aperture time of GEO SAR distinguishes it from the processing methods of low-orbit satellites with short synthetic aperture time in terms of target imaging, target detection, and target parameter estimation. Under long synthetic aperture time, ship targets are affected by sea surface motion and their own complex motion, and cannot be coherently accumulated for a long time on the ocean, resulting in difficulty in focusing and imaging of ship targets on the sea surface. In turn, when detecting ship targets, the signal-to-noise ratio of GEO SAR images is relatively low. Summary of the invention

[0003] The technical problems to be solved by the present invention are:

[0004] In order to avoid the shortcomings of the prior art, the present invention provides a method for ship moving target detection and track generation based on GEO SAR image sequence, which is used to solve the problems that ship targets cannot be coherently accumulated for a long time and target sequence detection is difficult under ultra-long synthetic aperture time of GEO SAR system.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A method for detecting moving targets and generating tracks of ships based on GEO SAR image sequences, characterized by comprising:

[0007] Acquire GEO SAR ship imaging data, divide the full aperture time into a number of sub-aperture times; image the GEO SAR ship imaging data respectively according to the number of sub-aperture times to obtain a sub-aperture image sequence;

[0008] Performing K-distribution CFAR detection on the sub-aperture image sequence to obtain a coarse target detection result;

[0009] Based on the coarse target detection result, calculating the centroid position of the sub-aperture image sequence;

[0010] Based on the centroid position of the sub-aperture image sequence, obtaining the ship track and estimating motion parameters; the motion parameters include range velocity and azimuth velocity;

[0011] Based on the motion parameters, the GEO SAR ship imaging data is subjected to secondary coherent accumulation to obtain a final ship moving target focused imaging result.

[0012] A further technical solution of the present invention: after obtaining the sub-aperture image sequence and before performing K-distribution CFAR detection, it also includes:

[0013] When the imaging of the sub-aperture image sequence does not adopt the BP imaging method, a certain sub-aperture image in the sub-aperture image sequence is selected as a reference sub-aperture image, and other sub-aperture images are registered to the reference sub-aperture image.

[0014] A further technical solution of the present invention is to perform K-distribution CFAR detection on the sub-aperture image sequence to obtain a coarse target detection result; comprising:

[0015] extracting the amplitude of the sub-aperture image sequence as sample data;

[0016] Determine the maximum amplitude and the minimum amplitude in the sample data, and determine an initial threshold based on the maximum amplitude and the minimum amplitude;

[0017] Based on the initial threshold, iteratively obtain the optimal threshold according to the K distribution bisection method;

[0018] Based on the sample data and the optimal threshold, a coarse target detection result is obtained.

[0019] A further technical solution of the present invention is to obtain a rough target detection result based on the sample data and the optimal threshold value; comprising:

[0020] Compare the amplitude of the sample data with the optimal threshold value. When the amplitude of the sample data is greater than the optimal threshold value, the pixel point corresponding to the amplitude is the target pixel point, and the value is 1; when the amplitude of the sample data is less than the optimal threshold value, the pixel point corresponding to the amplitude is the background clutter pixel point, and the value is 0;

[0021] The target pixels and the background clutter pixels are aggregated into a detection matrix, which is a coarse target detection result.

[0022] A further technical solution of the present invention is to calculate the centroid position of the sub-aperture image sequence based on the coarse target detection result; comprising:

[0023] Acquire the coordinates of the pixel points with a value of 1 in the rough target detection result in the target main area as the first coordinates, and calculate the second coordinates based on all the first coordinates and corresponding weights;

[0024] The corresponding weights obtained are summed to obtain a total weight value;

[0025] The coordinates (x, y) of the center of mass are calculated based on the abscissa, ordinate and the total weight value of the second coordinate, wherein Among them, X is the horizontal coordinate, Y is the vertical coordinate, and ρ is the total weight value.

[0026] A further technical solution of the present invention is to estimate motion parameters based on the centroid position of the sub-aperture image sequence; comprising:

[0027] Performing straight line fitting on the centroid position of the sub-aperture image sequence to obtain a horizontal coordinate difference and a vertical coordinate difference of the fitted straight line;

[0028] The range velocity is calculated based on the abscissa difference, the resolution m×n of the imaging grid and the synthetic aperture time: Among them, Δx is the difference between the horizontal and vertical coordinates, T on is the synthetic aperture time;

[0029] The azimuth velocity is calculated based on the ordinate difference, the resolution m×n of the imaging grid and the synthetic aperture time: Among them, Δx is the difference between the horizontal and vertical coordinates, T on is the synthetic aperture time.

[0030] A further technical solution of the present invention is to perform secondary coherent accumulation on the GEO SAR ship imaging data based on the motion parameters to obtain a final ship moving target focused imaging result; comprising:

[0031] Performing range-wise FFT on the GEO SAR ship imaging data to the range frequency domain;

[0032] Obtaining a radial velocity based on the range velocity, constructing a first matched filter function based on the radial velocity, and using the first matched filter function to perform range movement correction on the GEO SAR ship imaging data after range FFT processing;

[0033] Constructing a second matched filter function based on the azimuth velocity, and using the second matched filter function to perform distance frequency domain filtering on the distance movement data;

[0034] The data after the range frequency domain filtering is transformed into the range time domain-azimuth time domain by IFFT in the range direction, and the range frequency domain-azimuth frequency domain data is imaged by using the BP imaging algorithm to obtain the final ship moving target imaging result.

[0035] A computer system, characterized in that it comprises: one or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned method.

[0036] A computer-readable storage medium, characterized in that it stores computer-executable instructions, which are used to implement the above method when executed.

[0037] A computer program product, characterized by comprising computer executable instructions, wherein the instructions are used to implement the above method when executed.

[0038] The beneficial effects of the present invention are:

[0039] The present invention provides a method for ship target detection and track generation of GEO SAR image sequence, which is used for sequence detection, track generation and secondary coherent accumulation imaging of ship targets under long synthetic aperture time of GEO SAR system. Sub-aperture image sequence is generated by dividing long time into separate imaging, ship target detection is realized by using K-distribution CFAR detection, the weighted centroid estimation method is used to obtain the centroid position of the ship in the image sequence, a straight line is fitted to it and motion parameters are estimated, and secondary coherent accumulation of the image is realized according to the motion parameter estimation result. The present invention solves the problem that ship targets cannot be coherently accumulated for a long time and target sequence detection is difficult under ultra-long synthetic aperture time of GEO SAR system, and improves the image focusing accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0041] Figure 1 is a flow chart of the present invention;

[0042] Figure 2 This is the K distribution CFAR detection process in the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0044] The present invention proposes a method for ship moving target detection and track generation of GEO SAR image sequence, which divides a long synthetic aperture into sub-apertures and separately images to obtain the GEO SAR image sequence, performs target detection and motion parameter estimation on the image sequence to realize secondary coherent accumulation.

[0045] The technical idea of ​​implementing the present invention is: first, the GEO SAR original echo is input, and the sub-aperture data is obtained by sliding window method and imaged to generate an image sequence. Taking into account the time difference of each sub-aperture, the azimuth offset between each sub-aperture image can be estimated by using the land static scene, and each sub-aperture image is aligned before subsequent rough detection. For preliminary ship moving target detection, the K distribution ship target detection method can be used. After the preliminary detection is completed, it is necessary to estimate the centroid position and linear fitting of the rough detection results to generate the approximate track of the ship, and estimate the motion parameters of the ship moving target, including the speed of the ship moving target in azimuth and range. Then, these motion parameters are used to perform secondary coherent accumulation on the target to achieve image refocusing and repositioning correction.

[0046] The steps implemented by the present invention include the following:

[0047] (1) Ship moving target image sequence generation and image registration technology;

[0048] (1a) dividing the synthetic aperture time into sub-aperture times and imaging them separately to obtain an image sequence;

[0049] (1b) performing image registration on the image sequence data;

[0050] (2) K-distributed CFAR ship target detection;

[0051] (2a) using image sequence data as sample data;

[0052] (2b) Set the false alarm probability P fa ;

[0053] (2c) Setting the approximation accuracy σ;

[0054] (2d) The maximum amplitude is a and the minimum amplitude is b according to the amplitude of the sample data, and the loop variable n is set to 0, and the maximum number of loops is set to N;

[0055] (2e) Obtain the optimal threshold T according to the K distribution bisection method;

[0056] (2f) Compare the sample data pixel by pixel with the threshold T to obtain the detection result;

[0057] (3) Obtaining weighted centroid in image domain;

[0058] (3a) Using sliding window operation to select the target concentration area in the target detection results;

[0059] (3b) Setting the ratio of the target pixel in the window to the total pixel in the window;

[0060] (3c) The weight ρ in the centroid estimation iSet to the amplitude of the point in the image domain;

[0061] (3d) Accumulate the coordinates of the target pixel in the window and the weight of the point, divide the two and get the coordinate position (x, y) of the center of mass of the image;

[0062] (3e) Obtain and record the centroid coordinates of the image sequence;

[0063] (4) Ship track generation and motion parameter estimation;

[0064] (4a) Place the coordinate positions of the centroid of the image sequence in the same coordinate system and perform a straight line fitting on them;

[0065] (4b) According to the horizontal and vertical coordinate differences of the fitting line, the imaging grid size m×n and the synthetic aperture time T on , we can get the range velocity and azimuth velocity;

[0066] (4c) Connecting the centroid positions of the image sequence can obtain the approximate track of the ship;

[0067] (5) Secondary coherent accumulation of ship targets;

[0068] (5a) Performing range FFT transformation on the post-pulse compression data to the range frequency domain;

[0069] (5b) The radial velocity is obtained based on the estimated value of the range velocity, and the frequency domain filter is multiplied in the range frequency domain-azimuth time domain. Complete distance travel correction;

[0070] (5c) Based on the estimated azimuth velocity, multiply the frequency domain filter in the range frequency domain-azimuth time domain Performing phase compensation to achieve secondary coherent accumulation of image sequences;

[0071] (5d) performing range IFFT on the matched filtered data to transform it into the range time domain-azimuth time domain, and using the BP imaging algorithm to obtain the imaging result of the moving ship target;

[0072] (6) According to the above steps (1) to (5), the image sequence is iterated multiple times to estimate the ship motion parameters and perform phase compensation to achieve a better refocusing effect.

[0073] The specific implementation of the above steps is as follows:

[0074] Step 1: Ship target sub-aperture image sequence generation and image registration technology

[0075] (1a) The full-aperture imaging time of GEO SAR is above 100 seconds. It is possible to consider obtaining a scene image sequence based on the sub-aperture sliding window processing mode, dividing the full-aperture time into several sub-aperture times and imaging them separately. There should be a certain amount of data overlap between the previous and next sub-aperture images, and the higher the overlap rate, the shorter the time interval for image update. The imaging method of this step can adopt the BP imaging algorithm.

[0076] (1b) After the sub-aperture image sequence is generated, it is necessary to select a certain sub-aperture image as a reference and align the other sub-aperture images to the reference sub-aperture image. Alternatively, when the BP imaging method is selected in step 1a, if all sub-aperture images use the same imaging grid, there is no need to perform sub-aperture image registration.

[0077] Step 2: Perform K-distribution CFAR detection on the sub-aperture image sequence

[0078] (2a) The amplitude of the sub-aperture SAR image sequence data is taken as the sample data.

[0079] (2b) The amplitude of isolated strong scattering points of background clutter is greater than that of weak scattering points of the target. During detection, the strong scattering points of background clutter will be detected as target pixels. The false alarm probability P can usually be set according to the characteristics of the sample data. fa , set the false alarm probability to a small value, the false alarm situation is improved, usually set to 10 -3 ~10 -5 The false alarm probability set in the experiment is 10 -3 .

[0080] (2c) The setting of approximation accuracy σ in the binary method for thresholding. The smaller the setting, the closer the threshold is to the real threshold and the better the detection effect. However, if the setting is small, the amount of calculation will increase. Under the condition of a certain number of cycles, the threshold may not be obtained. Therefore, the setting accuracy σ should be set according to the sample characteristics and can generally be set to 10. -10 ~10 -15 , the value set in the experiment is 10 -15 .

[0081] (2d) The maximum amplitude is a and the minimum amplitude is b through the amplitude of the sample data, and the loop variable n is set to 0, and the maximum number of loops is set to N. The larger the number is, the greater the amount of calculation becomes, so it needs to be set appropriately. The experiment is set to 1000 times.

[0082] (2e) Derivation of K distribution threshold: The statistical model of sea clutter is set to K distribution, and the K distribution CFAR detector is used to detect ship targets, where the probability density function of K distribution is p(x): Where K v(x) is the v-order second-class modified Bessel function, Γ(x) is the Gamma function. c is the scale parameter, and v is the shape parameter. These two parameters are estimated in the statistical modeling process of K distribution;

[0083] Since the expression of p(x) is relatively complex, it is not easy to solve and integrate when deriving the threshold expression. The binary method can be used to solve the threshold T. The specific steps of this method are as follows:

[0084] make Where x is the amplitude of the sub-aperture image pixel, then:

[0085] 1. Let the minimum amplitude value be a=min(x), the maximum amplitude value be b=max(x), the loop variable n=0, and the maximum number of loops N and the approximation accuracy σ be given;

[0086] 2. Order It can be determined by F(T0), if |F(T0)-(1-P fa )|≤σ, go directly to step 4, otherwise go to step 3;

[0087] 3. If n>N, go to step 4; otherwise, when F(T0)<1-P fa When a=σ, when F(T0)>1-P fa When , let b = σ and go to step 2;

[0088] 4. Set T = σ and exit.

[0089] At the same time, when detecting different sub-aperture image sequences, the threshold T of a certain image in the image sequence can be adaptively determined according to the data features of the image, and the threshold is used to realize ship detection in the image, and the threshold is not applicable to other images.

[0090] (2f) Set the target pixel matrix det with the same size as the sample data. According to the threshold T, traverse all the pixel points of the data, and compare the amplitude of the pixel point with the threshold T one by one. The pixel point with an amplitude greater than the threshold T is determined to be a target pixel point, and the corresponding position of the det matrix is ​​written as 1; on the contrary, the pixel point with an amplitude less than the threshold T is determined to be a background clutter pixel point, and the corresponding position of the det matrix is ​​written as 0. Traverse the sample matrix, and the target pixel matrix det can be obtained, which is recorded as the detection matrix, and the target detection result is obtained.

[0091] Step 3: Image domain weighted centroid estimation

[0092] (3a) According to the ship detection results, the main target area is selected. When calculating the centroid position, the area where the target pixels are relatively concentrated is needed to increase the accuracy of the centroid estimation. Therefore, a sliding window operation is used to select the main target area. In the experiment, the window can be set according to the size of the target area, which can be set to [159,29], where 159 corresponds to the step length of the row and 29 corresponds to the step length of the column.

[0093] (3b) Set the window threshold ratio, that is, the ratio of the target pixel points in the window to the total pixel points in the window. In the experiment, the detection results are stored in the det matrix, where the target pixel point is 1 and the background clutter pixel point is 0. Therefore, it is only necessary to obtain the number of 1s in the det matrix in the window and divide it by the total number of pixels in the window (160*30), that is, the window ratio. At the same time, set the threshold ratio to 0.3. Slide the window from the initial position [1,1], calculate the corresponding window ratio pixel by pixel, and compare the size of the window ratio and the threshold ratio. When the window ratio is less than the threshold ratio, it means that the current window is not the main target area, and continue to slide the window; when the window ratio is greater than the threshold ratio, it means that the current window has found the main target area, stop sliding the window, and record the coordinates of the first point in the current window (x0, y0) to proceed to the next step. The size of the threshold ratio can be flexibly set. When the target pixels are sparse, the threshold ratio can be set smaller, and when the target pixels are relatively concentrated, it can be set higher.

[0094] (3c) Assuming that the ship target has uniform characteristics, it means that the target's centroid coincides with its geometric center. According to the centroid calculation formula: Among them, X0 is the horizontal coordinate of the center of mass of the ship target, N is the total number of pixels contained in the ship target, and x i is the horizontal coordinate of the i-th pixel of the ship target, ρ i is the mass density or amplitude weight of the i-th pixel of the ship target. Since the true mass density of each pixel of the ship target cannot be known exactly, it is assumed that the target has uniform characteristics, and the above formula can be simplified to: Similarly, the ordinate of the corresponding ship target center of mass can be obtained as:

[0095] Since the actual mass density ρ of the sub-aperture SAR image i Unknown, unweighted average centroid estimation assumes that the ship target has uniform characteristics, so that ρ i =1. However, in reality, there are strong and weak scattering points of ships. The amplitude value of the strong scattering points of ship targets in SAR images is greater than the amplitude value of the weak scattering points, so the entire ship target is not evenly distributed. Therefore, the centroid can be estimated based on the amplitude of the target pixel points in the SAR image, that is, for different pixel points, the weight value of the point can be set to iIt can be equal to the amplitude value of the point in the SAR image to estimate the centroid.

[0096] (3d) Select the main target area, start from the first point coordinate in the window, traverse all points in the window, determine the value of the corresponding coordinate point in the det matrix, and change its coordinate (x i ,y i ) and record all x values ​​that are 1 i and i Multiply by the weight ρ i , add up all the coordinates to get and And record it, and at the same time set the ρ of the point with value 1 i Accumulation Where N is the total number of points with a value of 1.

[0097] Therefore, we can get the values ​​of the horizontal coordinate X, the vertical coordinate Y, and the weight ρ, and then calculate the coordinates (x, y) of the center of mass, where The x and y coordinates are the horizontal and vertical coordinates of the center of mass. Since the coordinates are integer values, you need to round down x and y to get integer values.

[0098] (3e) Perform steps 3a-3e on all sub-aperture image sequence detection results to obtain the centroid position coordinates (x, y) of each sub-aperture and record them.

[0099] Step 4, ship track generation and motion parameter estimation;

[0100] (4a) Place the centroid of the image sequence in the same coordinate system, x i Represents the horizontal coordinate (distance direction) of the center of mass, y i represents the ordinate (azimuth) of the centroid, where i = 1 to N (N is the number of subapertures); a straight line is fitted to the centroid position, X i represents the horizontal coordinate of the fitted line, Y i Represents the ordinate of the fitted line, where i = 1 to N; Δx = X N -X1 represents the distance position difference after trajectory fitting, Δy=Y N -Y1 represents the position difference after trajectory fitting;

[0101] (4b) According to the horizontal and vertical coordinate differences Δx and Δy of the fitting line, the BP imaging grid size is m×n, and the synthetic aperture time T on ; Therefore, the estimated range velocity and azimuth velocity are: and

[0102] (4c) Based on the ship's centroid position result, the ship's centroids in all sub-aperture image sequences can be connected to obtain the ship's approximate track;

[0103] Step 5: Secondary coherent accumulation of ship targets

[0104] (5a) Performing range FFT on the echo data after pulse compression to the range frequency domain, the pulse compression echo signal is converted into the range frequency domain-azimuth time domain;

[0105] (5b) Using the distance velocity V estimated in step 4 x , and the radial velocity V is obtained r , in the range frequency domain-azimuth time domain, multiplied by the matched filter function where f c is the signal carrier frequency, c is the speed of light, f r represents the distance frequency, t a It is the azimuth moment to complete the distance movement correction of the moving target.

[0106] (5c) Using the azimuth velocity V estimated in step 4 y , in the range frequency domain-azimuth time domain, multiplied by the matched filter function, Where V a is the flight speed of the radar platform, R0 is the closest distance between the target and the radar, and the secondary coherent accumulation of the ship target is completed in the range time domain-azimuth time domain.

[0107] (5d) performing range IFFT transformation on the matched filtered data to the range time domain-azimuth time domain, and using the BP imaging algorithm to obtain the final ship moving target imaging result;

[0108] In step 6, steps 1 to 4 may be used to perform multiple ship motion parameter estimation and phase compensation on the GEO SAR image sequence to achieve a better focusing effect.

[0109] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should be included in the protection scope of the present invention.

Claims

1. A method for detecting moving targets and generating tracks of ships based on GEO SAR image sequences, characterized in that: include: Acquire GEO SAR ship imaging data and divide the full aperture time into several sub-aperture times; The GEO SAR ship imaging data is imaged respectively according to the plurality of sub-aperture times to obtain a sub-aperture image sequence; Performing K-distribution CFAR detection on the sub-aperture image sequence to obtain a coarse target detection result; Based on the coarse target detection result, calculating the centroid position of the sub-aperture image sequence; Based on the centroid position of the sub-aperture image sequence, obtaining the ship track and estimating motion parameters; the motion parameters include range velocity and azimuth velocity; Based on the motion parameters, the GEO SAR ship imaging data is subjected to secondary coherent accumulation to obtain a final ship moving target focused imaging result.

2. The method for detecting moving targets and generating tracks of ships based on GEO SAR image sequences according to claim 1, characterized in that: After obtaining the sub-aperture image sequence and before performing K-distribution CFAR detection, it also includes: When the imaging of the sub-aperture image sequence does not adopt the BP imaging method, a certain sub-aperture image in the sub-aperture image sequence is selected as a reference sub-aperture image, and other sub-aperture images are registered to the reference sub-aperture image.

3. The method for ship moving target detection and track generation based on GEO SAR image sequence according to claim 1, characterized in that: Performing K-distribution CFAR detection on the sub-aperture image sequence to obtain a coarse target detection result; comprising: extracting the amplitude of the sub-aperture image sequence as sample data; Determine the maximum amplitude and the minimum amplitude in the sample data, and determine an initial threshold based on the maximum amplitude and the minimum amplitude; Based on the initial threshold, iteratively obtain the optimal threshold according to the K distribution bisection method; Based on the sample data and the optimal threshold, a coarse target detection result is obtained.

4. The method for detecting moving targets and generating tracks of ships based on GEO SAR image sequences according to claim 3, characterized in that: Based on the sample data and the optimal threshold, a rough target detection result is obtained; comprising: Compare the amplitude of the sample data with the optimal threshold value. When the amplitude of the sample data is greater than the optimal threshold value, the pixel point corresponding to the amplitude is the target pixel point, and the value is 1; when the amplitude of the sample data is less than the optimal threshold value, the pixel point corresponding to the amplitude is the background clutter pixel point, and the value is 0; The target pixels and the background clutter pixels are aggregated into a detection matrix, which is a coarse target detection result.

5. The method for ship moving target detection and track generation based on GEO SAR image sequence according to claim 4, characterized in that: Based on the coarse target detection result, calculating the centroid position of the sub-aperture image sequence; comprising: Acquire the coordinates of the pixel points with a value of 1 in the rough target detection result in the target main area as the first coordinates, and calculate the second coordinates based on all the first coordinates and corresponding weights; The corresponding weights obtained are summed to obtain a total weight value; The coordinates (x, y) of the center of mass are calculated based on the abscissa, ordinate and the total weight value of the second coordinate, wherein Among them, X is the horizontal coordinate, Y is the vertical coordinate, and ρ is the total weight value.

6. The method for ship moving target detection and track generation based on GEO SAR image sequence according to claim 1, characterized in that: Based on the centroid position of the sub-aperture image sequence, estimating motion parameters; comprising: Performing straight line fitting on the centroid position of the sub-aperture image sequence to obtain a horizontal coordinate difference and a vertical coordinate difference of the fitted straight line; The range velocity is calculated based on the abscissa difference, the resolution m×n of the imaging grid and the synthetic aperture time: Among them, Δx is the horizontal and vertical coordinate difference, Ton is the synthetic aperture time; The azimuth velocity is calculated based on the ordinate difference, the resolution m×n of the imaging grid and the synthetic aperture time: Wherein, Δx is the horizontal and vertical coordinate difference, and T is the synthetic aperture time.

7. The method for ship moving target detection and track generation based on GEO SAR image sequence according to claim 1, characterized in that: Based on the motion parameters, the GEO SAR ship imaging data is subjected to secondary coherent accumulation to obtain a final ship moving target focused imaging result; including: Performing range-wise FFT on the GEO SAR ship imaging data to the range frequency domain; Obtaining a radial velocity based on the range velocity, constructing a first matched filter function based on the radial velocity, and using the first matched filter function to perform range movement correction on the GEO SAR ship imaging data after range FFT processing; Constructing a second matched filter function based on the azimuth velocity, and using the second matched filter function to perform distance frequency domain filtering on the distance movement data; The data after the range frequency domain filtering is transformed into the range time domain-azimuth time domain by IFFT in the range direction, and the range frequency domain-azimuth frequency domain data is imaged by using the BP imaging algorithm to obtain the final ship moving target imaging result.

8. A computer system, characterized in that include: One or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method of claim 1.

9. A computer-readable storage medium, characterized in that Computer executable instructions are stored, and when the instructions are executed, they are used to implement the method of claim 1.

10. A computer program product, characterized in that The invention comprises computer executable instructions, which are used to implement the method of claim 1 when being executed.

Images