Multi-angle SAR moving target imaging method and device based on compressed sensing algorithm

Abstract

This invention provides a multi-angle SAR moving target imaging method and apparatus based on compressed sensing algorithms, comprising: dividing the original radar echo signal into sub-apertures to obtain multi-channel original echo signals; preprocessing the multi-channel original echo signals in the range-Doppler domain to construct a moving target imaging model; correcting translational motion errors using a Legendre polynomial fitting method; performing sparse reconstruction using the alternating direction multiplier method to obtain the moving target imaging result in the range-Doppler domain, and introducing a two-dimensional spatially varied error matrix into the alternating direction multiplier method algorithm framework to compensate for two-dimensional spatially varied phase errors; and projecting the moving target imaging result in the range-Doppler domain using two-dimensional interpolation based on the relationship between the range-Doppler domain and the northeast coordinate system to obtain the moving target focused imaging result in the ground-level plane. This invention can effectively achieve moving target refocusing imaging, azimuth calibration, and attitude correction.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and in particular to a multi-angle SAR moving target imaging method and apparatus based on compressed sensing algorithms. Background Technology

[0002] Synthetic aperture radar ground moving target indication (SAR-GMTI) technology has become a major research focus in the SAR field in recent years. With the deepening of research, its application scope is constantly expanding. In military applications, it is used for battlefield reconnaissance and intelligence gathering; in civilian applications, it is used for traffic control under adverse weather / nighttime conditions and information support for disaster relief. Wide-angle staring SAR (WasSAR) is a special SAR mode that achieves long-term observation of key areas by flying along curves or circles. Unlike traditional linear synthetic aperture radar (LSAR-GMTI), WasSAR-GMTI can achieve long-term continuous tracking and monitoring of moving targets. Therefore, WasSAR-GMTI can effectively compensate for the shortcomings of traditional linear SAR-GMTI and has a wider range of applications in both military and civilian fields. In WasSAR mode, the complex trajectory of the carrier aircraft leads to severe two-dimensional coupling between azimuth and range. Furthermore, the complex maneuvers of non-cooperative targets make traditional imaging methods difficult to apply to moving target imaging in WasSAR mode. Therefore, moving target refocusing imaging in WasSAR mode remains a challenge in the GMTI field. Ground moving target (GMT) imaging can generally be divided into parametric and non-parametric imaging methods. Parametric imaging methods involve establishing an accurate SAR moving target imaging model, estimating the moving target's velocity parameters or Doppler parameters (center and frequency modulation), and then designing a matching function to compensate for range migration and azimuth defocus caused by motion. However, most parametric imaging methods are based on LSAR mode, and the unique motion geometry of WasSAR introduces severe azimuth and range coupling. This makes these methods difficult to apply to WasSAR mode. Non-parametric imaging methods transform the moving target imaging problem into an optimization problem. Methods based on Inverse Synthetic Aperture (ISAR) are an important class of non-parametric imaging methods. ISAR technology can achieve high-precision imaging of non-cooperative moving targets. Its imaging processing does not depend on the estimation results of target motion parameters, giving it an inherent advantage in applying ISAR technology to WasSAR ground moving target imaging. In recent years, some researchers have combined ISAR imaging with sparse reconstruction algorithms, reconstructing high-resolution ISAR images by solving a sparsity-driven optimization problem, and compensating for higher-order phase errors using the low-rank nature of vibration errors or based on image minimum entropy. However, existing methods lack two-dimensional spatially varying error compensation methods for complex maneuvering targets. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a multi-angle SAR moving target imaging method and apparatus based on compressed sensing algorithms.

[0004] To achieve the above technical objectives, the specific technical solution adopted by the present invention is as follows:

[0005] On the one hand, the present invention provides a multi-angle SAR moving target imaging method based on compressed sensing algorithm, comprising the following steps:

[0006] The original radar echo signal is divided into sub-apertures to obtain multi-channel original echo signals;

[0007] The multi-channel raw echo signal is preprocessed in the range Doppler domain, including: multi-channel registration, clutter suppression and radial velocity estimation; an imaging model is constructed on the moving target echo signal obtained after clutter suppression, the moving target imaging model includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated;

[0008] A method based on Legendre polynomial fitting was used to correct translational motion errors.

[0009] Sparse reconstruction is performed using the alternating direction multiplier method to obtain the moving target imaging results in the range-Doppler domain. The two-dimensional spatially variable error matrix is ​​introduced into the alternating direction multiplier method algorithm framework to compensate for the two-dimensional spatially variable phase error.

[0010] Based on the relationship between the range-Doppler domain and the northeast coordinate system, a two-dimensional interpolation method is used to project the moving target imaging results in the range-Doppler domain to obtain the moving target focused imaging results in the ground-range plane.

[0011] Furthermore, the moving target imaging model is constructed according to the following steps:

[0012] Range pulse compression and range-directed fast Fourier transform are performed on the multi-channel raw echo signals to obtain the target's echo fundamental frequency signal.

[0013]

[0014] in, , and These represent the bandwidth, carrier frequency, and speed of light of the echo signal, respectively. It represents fast time. Represents slow time; Represents the carrier aircraft to the The slant distance of each scattering center This indicates the total number of scattering points contained in the moving target; j The imaginary unit;

[0015] The target motion is transformed into an ideal turntable model according to the following formula.

[0016]

[0017] in, In order to be in The instantaneous change in slant distance caused by rotation at any given moment; and Let x and y be the x and y coordinates of the p-th scattering center; The distance between the reference point and the aircraft; The rotation angle of the carrier aircraft around the ideal circular arc with the target center as the origin on the LOS plane; The rotation angle of the moving target around the origin;

[0018] The carrier aircraft arrived at the Slope distance of each scattering center Expressed as:

[0019]

[0020] in, and These represent the radial velocity of the target and the radial velocity of the carrier aircraft, respectively. Indicates translational motion;

[0021] Will According to Taylor expansion, we get:

[0022]

[0023] in, , These represent the first and second spatial variations of distance and azimuth, respectively. Order coefficient; Represents any time Projection onto the reference imaging projection plane;

[0024] After Taylor was unfolded The moving target imaging model is obtained by incorporating the target's echo fundamental frequency signal:

[0025] .

[0026] Furthermore, the multi-channel raw echo signal undergoes preprocessing in the range Doppler domain, including the following steps:

[0027] Azimuth compression is achieved by performing a fast Fourier transform on the multi-channel raw echo signal in the azimuth dimension to obtain the range Doppler echo;

[0028] A multi-channel moving target detection algorithm is used to process the range Doppler echo, and clutter-suppressed moving target echo signal and target radial velocity estimation results are obtained.

[0029] Furthermore, the method for correcting translational motion errors using Legendre polynomial fitting includes the following steps:

[0030] Extracting the moving target echo phase from the moving target imaging model

[0031]

[0032] in, Represents phase extraction; Represents phase along Untangling; This indicates a fast Fourier transform along the fast time axis;

[0033] Phase error fitting using Legendre polynomials

[0034]

[0035] in, Represents the number of samples along the distance direction; Representing Legendre polynomials, Represents the coefficients of a polynomial; This represents the estimated residual translational error; For distance to the first Each sampling frequency point The sampling frequency is the range direction.

[0036] Construct the compensation function:

[0037] ;

[0038] The translational motion error is compensated using the following formula, resulting in the moving target imaging model after translational motion error compensation:

[0039]

[0040] in, This indicates the inverse fast Fourier transform along the fast time frequency.

[0041] Furthermore, the method of using alternating direction multipliers for sparse reconstruction to obtain the moving target imaging results in the range-Doppler domain includes the following steps:

[0042] The moving target imaging model after translational motion error compensation is divided into blocks along the azimuth dimension in the range-Doppler domain.

[0043]

[0044]

[0045]

[0046] in, Represents pulse duration; Represents convolution operation; Represents the Doppler frequency; This represents the phase error in the range-Doppler domain; It is a window function along the orientation dimension. Represents the azimuth sampling frequency. Represents the number of blocks. Represents the Doppler frequency interval; Represents the number of blocks; Represents the number of sampling points in the azimuth direction;

[0047] right Inverse Fast Fourier Transform along the azimuth dimension yields the block data in the range-azimuth-time domain. ;

[0048] Using the alternating direction multiplier method, for The solution is performed to obtain well-focused segmented sub-images;

[0049] The moving target imaging results in the range-Doppler domain are obtained by summing up all the segmented sub-images.

[0050] Furthermore, the method of alternating direction multipliers is employed for... The solution process yields well-focused sub-image blocks, including the following steps:

[0051] The solution process is broken down into the following three parts, where the first part is... m The first position sub-block k +1 iterations can be represented as:

[0052]

[0053] in, express The target image can be represented in matrix form as... ; Represents the Fourier matrix along the azimuth direction; Represents conjugate transpose; Represents auxiliary variables; Represents the dual variable; Represents the regularization coefficient; Represents a threshold function; Represents the spatially variable phase compensation matrix;

[0054] In the formula Expressed as

[0055]

[0056] in, Representing the Each distance unit, with phase compensation factor along the azimuth dimension;

[0057] The phase compensation factor for each range cell is solved using the following formula.

[0058]

[0059]

[0060] in, Representing the n An image of a distance unit; Representing the image number n The distance unit below the first i Values ​​of each azimuth unit; Indicates the first n Echoes from each distance cell, Indicates the first i The Fourier matrix corresponding to each square Doppler frequency value;

[0061] The phase compensation factor is processed using the image sharpest criterion.

[0062]

[0063] in, Indicates the sharpness of an image;

[0064] The phase compensation factor for each distance cell is solved using the image sharpest criterion.

[0065] By fitting the data along the range direction, the estimated phase error values ​​corresponding to all range cells are obtained. ;

[0066] Will Incorporating the alternating direction multiplier algorithm framework In the process of correcting the two-dimensional spatial phase error, after multiple iterations using the alternating direction multiplier method, a well-focused segmented sub-image is obtained.

[0067] Furthermore, the moving target imaging result in the range-Doppler domain is expressed as follows:

[0068] .

[0069] Furthermore, the step of projecting the moving target imaging result in the range-Doppler domain onto the ground-based coordinate system using two-dimensional interpolation based on the relationship between the range-Doppler domain and the northeast coordinate system to obtain the moving target focused imaging result in the ground-based distance plane includes the following steps:

[0070] Based on the target radial velocity estimate Construct Doppler offset values

[0071] ;

[0072] Translate the image along the Doppler direction

[0073] ;

[0074] Based on the relationship between the distance Doppler domain and the northeast coordinate system, two-dimensional interpolation is used to... Projected to This yields the focused imaging results of the moving target at the ground distance plane.

[0075] Furthermore, the relationship between the distance Doppler domain and the northeast coordinate system is as follows:

[0076]

[0077] in, , , , These are the aircraft's location and speed information recorded by GPS.

[0078] On the other hand, the present invention provides a multi-angle SAR moving target imaging device based on compressed sensing algorithm, comprising:

[0079] The first module is used to divide the original radar echo signal into sub-apertures to obtain multi-channel original echo signals.

[0080] The second module is used to preprocess the multi-channel raw echo signal in the range Doppler domain, including: multi-channel registration, clutter suppression and radial velocity estimation; and to construct an imaging model for the moving target echo signal obtained after clutter suppression. The moving target imaging model includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated.

[0081] The third module is used to correct translational motion errors using a method based on Legendre polynomial fitting.

[0082] The fourth module is used to perform sparse reconstruction using the alternating direction multiplier method to obtain the moving target imaging results in the range-Doppler domain. It introduces the two-dimensional spatially variable error matrix into the alternating direction multiplier method algorithm framework to compensate for the two-dimensional spatially variable phase error.

[0083] The fifth module is used to project the moving target imaging results in the range-Doppler domain using two-dimensional interpolation based on the relationship between the range-Doppler domain and the northeast coordinate system, so as to obtain the moving target focused imaging results in the ground-range plane.

[0084] Compared with the prior art, the present invention has the following beneficial technical effects:

[0085] The multi-angle SAR moving target imaging method and apparatus based on compressed sensing algorithm provided by this invention establishes a moving target imaging model that includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated, and uses a Legendre polynomial fitting method to correct the translational motion error. Then, the moving target imaging model after translational motion error compensation is divided into blocks along the azimuth dimension in the range-Doppler domain. For each block sub-image, the sparsity of the moving target is utilized, and sparse reconstruction is performed by combining alternating direction multiplication. During the reconstruction process, a vacuum-variable phase compensation matrix is ​​introduced, thereby correcting the two-dimensional vacuum-variable phase error during the alternating direction multiplication iteration process. This makes the image signal-to-noise ratio gradually increase with the number of iterations, and the residual vacuum-variable phase error gradually decrease, finally obtaining well-focused block sub-images.

[0086] Considering that the effective rotational angular velocity of a maneuvering target is time-varying, this invention utilizes the correspondence between the range-Doppler domain and the northeast-sky coordinate system, and uses two-dimensional interpolation to project the moving target imaging results in the range-Doppler domain, thereby solving the orientation calibration problem and correcting the moving target image to the actual ground plane attitude. Attached Figure Description

[0087] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0088] Figure 1 Here is a flowchart of a multi-angle SAR moving target imaging method based on compressed sensing algorithm provided in one embodiment;

[0089] Figure 2 This is a schematic diagram of the geometry for WasSAR moving target imaging provided in one embodiment;

[0090] Figure 3 This is a schematic diagram illustrating the equivalent of WasSAR moving target imaging and ISAR imaging in one embodiment.

[0091] Figure 4 An example of an experimental scene optical image and SAR imaging results is provided.

[0092] Figure 5 This is a diagram showing the experimental carrier aircraft and its GPS recording results provided in one embodiment.

[0093] Figure 6 An optical image of a cooperative vehicle target provided in one embodiment;

[0094] Figure 7 The following are images of moving targets provided in one embodiment: Figures (a), (f), and (k) are images of targets 1-3 based on the ISAR imaging method; Figures (b), (g), and (l) are images of targets 1-3 based on the ISAR imaging method combined with autofocusing algorithms; Figures (c), (h), and (m) are images of targets 1-3 based on the alternating direction multiplier method imaging algorithm; Figures (d), (i), and (n) are images of targets 1-3 based on the autofocusing method of the alternating direction multiplier method; and Figures (e), (j), and (o) are images of targets 1-3 based on the multi-angle SAR moving target imaging method using compressed sensing algorithms.

[0095] Figure 8 This is a target calibration result diagram in azimuth 1-3 provided in one embodiment;

[0096] Figure 9 An imaging result diagram and corresponding optical image of three cooperating targets in each sub-aperture are provided for one embodiment. Detailed Implementation

[0097] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0098] Reference Figure 1 One embodiment provides a multi-angle SAR moving target imaging method based on compressed sensing algorithm, including the following steps:

[0099] The original radar echo signal is divided into sub-apertures to obtain multi-channel original echo signals;

[0100] The multi-channel raw echo signal is preprocessed in the range Doppler domain, including: multi-channel registration, clutter suppression and radial velocity estimation; an imaging model is constructed on the moving target echo signal obtained after clutter suppression, the moving target imaging model includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated;

[0101] A method based on Legendre polynomial fitting was used to correct translational motion errors.

[0102] Sparse reconstruction is performed using the alternating direction multiplier method to obtain the moving target imaging results in the range-Doppler domain. The two-dimensional spatially variable error matrix is ​​introduced into the alternating direction multiplier method algorithm framework to compensate for the two-dimensional spatially variable phase error.

[0103] Based on the relationship between the range-Doppler domain and the northeast coordinate system, a two-dimensional interpolation method is used to project the moving target imaging results in the range-Doppler domain to obtain the moving target focused imaging results in the ground-range plane.

[0104] Reference Figure 2 This is a schematic diagram of WasSAR moving target imaging geometry provided in one embodiment. In this embodiment, the SAR radar receives the original radar echo signal and performs sub-aperture division to obtain multi-channel original echo signals; range pulse compression and range-directed fast Fourier transform are performed on the multi-channel original echo signals to obtain the target's echo fundamental frequency signal.

[0105]

[0106] in, , and These represent the bandwidth, carrier frequency, and speed of light of the echo signal, respectively. It represents fast time. Represents slow time; Represents the carrier aircraft to the The slant distance of each scattering center This indicates the total number of scattering points contained in the moving target; j It is the imaginary unit.

[0107] The coherent accumulation angle formed by the aircraft flying along the curve relative to the target is the basis for high-resolution imaging in the WasSAR system. Similarly, the rotation of the target relative to the radar is also the essential source of ISAR imaging. Therefore, the imaging problem of moving targets in WasSAR can be solved by referring to the ISAR model. Existing ISAR models usually decompose the relationship between the radar and the target into translational motion and rotational motion. Translational motion refers to the effect on all scattering points on the target being the same, resulting in the same Doppler at different scattering points. Therefore, it does not contribute to imaging and will cause image shift and blurring. Without loss of generality, the target center is chosen as the reference point, and reference is made to... Figure 1 Translational motion It can be represented as

[0108]

[0109] in, and These represent the radial velocities of the target and the carrier aircraft, respectively. Assuming translational motion has been compensated, refer to... Figure 3 (a) On the LOS plane, with the target center as the origin, the carrier aircraft flies an ideal circular arc around the origin; the target motion in the WasSAR model is transformed into an ideal turntable model according to the following formula, such as Figure 3 As shown in (b):

[0110]

[0111] in, In order to be in The instantaneous change in slant distance caused by rotation at any given moment; and Let x and y be the x and y coordinates of the p-th scattering center; The distance between the reference point and the aircraft; The rotation angle of the carrier aircraft around the ideal circular arc with the target center as the origin on the LOS plane; The rotation angle of the moving target around the origin;

[0112] The carrier aircraft arrived at the Slope distance of each scattering center Expressed as:

[0113]

[0114] in, and These represent the radial velocity of the target and the radial velocity of the carrier aircraft, respectively. Indicates translational motion;

[0115] However, in reality, the goal is not to achieve ideal, stable motion, but rather to engage in maneuvering; maneuvering refers to motion that includes time-varying rotational speed and acceleration. Therefore, According to Taylor expansion, we get:

[0116]

[0117] in, , These represent the first and second spatial variations of distance and azimuth, respectively. Order coefficient; Represents any time Projection onto the reference imaging projection plane;

[0118] As can be seen from the above equation, range migration exists in the envelope term, requiring correction using an envelope alignment algorithm. The first phase term is a constant and can be ignored; the second phase term represents the translational phase error introduced by translational motion, which is not spatially variable; the third and fourth phase terms represent the range and azimuth spatially variable phase errors, respectively. It can be seen that, compared to traditional ISAR imaging, moving target imaging in WasSAR mode suffers from severe translational errors and two-dimensional spatially variable phase errors.

[0119] After Taylor was unfolded The moving target imaging model is obtained by incorporating the target's echo fundamental frequency signal:

[0120] ;

[0121] As can be seen from the above, the moving target imaging model includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated.

[0122] Azimuth compression is achieved by performing a Fast Fourier Transform (FFT) on the original multi-channel echo signals in the azimuth dimension, yielding range Doppler echoes. Due to the time delay differences caused by antenna spacing when observing the same scene from different channels, the correlation of the multi-channel received signals is disrupted; therefore, time delay error compensation is necessary. With increasing spatial degrees of freedom, the clutter suppression effect of multi-channel methods improves. Multi-channel moving target detection methods include phase-center offset antennas, track-along interferometry, clutter suppression, and space-time adaptive processing. Furthermore, the radial velocity of the moving target can be estimated using weighted adaptive matched filtering and subspace projection.

[0123] A multi-channel moving target detection algorithm is used to process the range Doppler echo, and clutter-suppressed moving target echo signal and target radial velocity estimation results are obtained.

[0124] The method for correcting translational motion errors based on Legendre polynomial fitting includes the following steps:

[0125] Extracting the moving target echo phase from the moving target imaging model

[0126]

[0127] in, Represents phase extraction; Represents phase along Untangling; This indicates a fast Fourier transform along the fast time axis;

[0128] Phase error fitting using Legendre polynomials

[0129]

[0130] in, Represents the number of samples along the distance direction; Representing Legendre polynomials, Represents the coefficients of a polynomial. This represents the estimated residual translational error; For distance to the first Each sampling frequency point The sampling frequency is the range direction.

[0131] Construct the compensation function:

[0132] ;

[0133] The translational motion error is compensated using the following formula, resulting in the moving target imaging model after translational motion error compensation:

[0134]

[0135] in, This indicates the inverse fast Fourier transform along the fast time frequency.

[0136] The method of using alternating direction multipliers for sparse reconstruction to obtain the moving target imaging results in the range-Doppler domain includes the following steps:

[0137] Because the phase error has spatially variable phase error along the azimuth direction, the moving target imaging model after translational motion error compensation is divided into blocks along the azimuth dimension in the range-Doppler domain. Specifically, for... Along Divide into blocks:

[0138]

[0139]

[0140]

[0141] in, Represents pulse duration; Represents convolution operation; Represents the Doppler frequency; This represents the phase error in the range-Doppler domain; It is a window function along the orientation dimension. Represents the azimuth sampling frequency. Represents the number of blocks. Represents the Doppler frequency interval; Represents the number of blocks; Represents the number of sampling points in the azimuth direction;

[0142] right Inverse Fast Fourier Transform along the azimuth dimension yields the block data in the range-azimuth-time domain. ;

[0143] Using the alternating direction multiplier method, for The solution process yields well-focused sub-image blocks, including the following steps:

[0144] The solution process is broken down into the following three parts, and to improve computational efficiency, the vector is converted into a matrix, where the first... m The first position sub-block k +1 iterations can be represented as:

[0145]

[0146] in, express The target image can be represented in matrix form as... ; Represents the Fourier matrix along the azimuth direction; Represents conjugate transpose; Represents auxiliary variables; Represents the dual variable; Represents the regularization coefficient; Represents a threshold function; Represents the spatially variable phase compensation matrix;

[0147] In the above formula, in order to preserve the structure of the moving target, the fixed part in the traditional method is... Change to Compared to the traditional alternating direction multiplier method, the above equation introduces a spatially variable phase compensation matrix, thereby reducing the two-dimensional spatially variable phase error.

[0148] In the formula Expressed as

[0149]

[0150] in, Representing the Each distance unit, with phase compensation factor along the azimuth dimension;

[0151] The phase compensation factor for each range cell is solved using the following formula.

[0152]

[0153]

[0154] in, Representing the n An image of a distance unit; Representing the image numbern The distance unit below the first i Values ​​of each azimuth unit; Indicates the first n Echoes from each distance cell, Indicates the first i The Fourier matrix corresponding to each square Doppler frequency value;

[0155] The phase compensation factor is processed using the image sharpest criterion.

[0156]

[0157] in, Indicates the sharpness of an image;

[0158] Using the image sharpest criterion, the phase compensation factor for each distance cell is solved; when After processing the data for each distance cell, perform data fitting along the distance direction to obtain the estimated phase error values ​​for all distance cells. ;

[0159] Will Incorporating the alternating direction multiplier algorithm framework In the process of correcting the two-dimensional spatial phase error, after multiple iterations using the alternating direction multiplier method, a well-focused segmented sub-image is obtained.

[0160] By using the alternating direction multiplier method for multiple iterations, the signal-to-noise ratio (SNR) of the image gradually increases (SNR = moving target signal energy / (residual stationary target signal energy + noise)), while the residual spatial phase error gradually decreases.

[0161] The moving target imaging result in the range-Doppler domain is obtained by summing all the segmented sub-images. The moving target imaging result in the range-Doppler domain is expressed as follows:

[0162] .

[0163] The method of projecting the moving target imaging results in the range-Doppler domain onto the ground-based coordinate system using two-dimensional interpolation based on the relationship between the range-Doppler domain and the northeast coordinate system to obtain the moving target focused imaging results in the ground-based distance plane includes the following steps:

[0164] Based on the target radial velocity estimate Construct Doppler offset values

[0165] ;

[0166] Translate the image along the Doppler direction

[0167] ;

[0168] Based on the relationship between the distance Doppler domain and the northeast coordinate system, two-dimensional interpolation is used to... Projected to This yields the focused imaging results of the moving target at the ground distance plane.

[0169] The relationship between the distance-Doppler domain and the northeast coordinate system is as follows:

[0170]

[0171] in, , , , These are the aircraft's location and speed information recorded by GPS.

[0172] In one embodiment, a Ku-band three-channel airborne WasSAR experiment was conducted to verify the effectiveness of the multi-angle SAR moving target imaging method based on compressed sensing algorithm.

[0173] The experimental scenario is a circular intersection. Optical and SAR images of the scenario are shown below. Figure 4 As shown. The radar system is installed on a CESSNA-172R (such as...). Figure 5 (a) shows the aircraft trajectory as follows Figure 5 As shown in (b). During the experiment, three pickup trucks were arranged as the experimental cooperation targets. Figure 4 (a) shows a portion of the tracks recorded by the vehicle's GPS. Figure 6 The exhibition showcases the pickup truck's front end, parking space, and body details. All three vehicles are identical in model and dimensions. The pickup truck has a relatively obvious dihedral structure at the front and rear, which facilitates subsequent verification of the size and attitude of moving targets. During the experiment, considering that a sub-aperture that is too small would result in low imaging resolution, while a sub-aperture that is too large would lead to complex movement of the moving target during the coherent accumulation time, the complete echo data was divided into sub-apertures of 2° each, with a sub-aperture overlap rate of 50%.

[0174] To verify the effectiveness of the method described in this invention, the multi-angle SAR moving target imaging method based on compressed sensing algorithm is compared with the imaging method based on ISAR, the imaging method based on ISAR combined with self-focusing algorithms, the imaging algorithm based on alternating direction multiplier method, and the self-focusing imaging method based on alternating direction multiplier method. Figure 7The images provided are moving target imaging results for one embodiment. Figures (a), (f), and (k) show the imaging results of targets 1-3 based on the ISAR imaging method; Figures (b), (g), and (l) show the imaging results of targets 1-3 based on the ISAR imaging method combined with autofocus algorithms; Figures (c), (h), and (m) show the imaging results of targets 1-3 based on the alternating direction multiplier method imaging algorithm; Figures (d), (i), and (n) show the imaging results of targets 1-3 based on the autofocus method of the alternating direction multiplier method; and Figures (e), (j), and (o) show the imaging results of targets 1-3 based on the multi-angle SAR moving target imaging method using compressed sensing algorithms. By comparison, it can be seen that the ISAR-based imaging method, besides compensating for translational errors, does not perform high-order phase compensation for rotational errors, resulting in severe defocusing along the azimuth direction. The ISAR-based imaging method combined with autofocus algorithms can effectively compensate for high-order phase errors of the target, but it suffers from low signal-to-noise ratio and unclear target contours. The alternating direction multiplier-based imaging algorithm can enhance images by utilizing the sparse nature of targets, thereby improving the signal-to-noise ratio. However, this method does not inherently estimate or compensate for phase errors, resulting in significant defocusing. While self-focusing imaging methods based on the alternating direction multiplier method estimate and compensate for phase errors, their focus quality is low due to the lack of consideration for the two-dimensional spatial variability of phase errors. The multi-angle SAR moving target imaging method based on compressed sensing algorithm provided in this invention combines the advantages of the alternating direction multiplier method for image enhancement with the introduction of a two-dimensional phase error compensation method for effective target focusing. Therefore, it can obtain moving target imaging results with high focus quality, and the attitude of the imaging result corresponds to the actual attitude of the moving target.

[0175] Because the imaging results are on the Los plane, they differ from the actual vehicle shape and attitude. The proposed method is used for orientation calibration and attitude correction, and the results are as follows: Figure 8 First, the size of the vehicle imaging result (shown). , as well as ) and the actual size of the pickup truck ( The images are largely consistent with the actual vehicle posture. Furthermore, the vehicle's posture is consistent with its actual posture. The imaging results show that the pickup truck's cab and cargo bed have dihedral structures, resulting in strong energy regions in the SAR image.

[0176] The proposed algorithm was used to refocus and image three cooperating targets in 200 frames of sub-aperture data. The imaging results are as follows: Figure 9As shown, the algorithm proposed in this paper can obtain moving target imaging results with high focusing quality, and the attitude of the moving target imaging results also corresponds to the actual attitude, proving the effectiveness of the algorithm in moving target refocusing imaging, orientation calibration, and attitude correction.

[0177] In another embodiment, a multi-angle SAR moving target imaging device based on compressed sensing algorithm is provided, comprising:

[0178] The first module is used to divide the original radar echo signal into sub-apertures to obtain multi-channel original echo signals.

[0179] The second module is used to preprocess the multi-channel raw echo signal in the range Doppler domain, including: multi-channel registration, clutter suppression and radial velocity estimation; and to construct an imaging model for the moving target echo signal obtained after clutter suppression. The moving target imaging model includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated.

[0180] The third module is used to correct translational motion errors using a method based on Legendre polynomial fitting.

[0181] The fourth module is used to perform sparse reconstruction using the alternating direction multiplier method to obtain the moving target imaging results in the range-Doppler domain. It introduces the two-dimensional spatially variable error matrix into the alternating direction multiplier method algorithm framework to compensate for the two-dimensional spatially variable phase error.

[0182] The fifth module is used to project the moving target imaging results in the range-Doppler domain using two-dimensional interpolation based on the relationship between the range-Doppler domain and the northeast coordinate system, so as to obtain the moving target focused imaging results in the ground-range plane.

[0183] Matters not covered in this invention are common knowledge.

[0184] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0185] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

[0186] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-angle SAR moving target imaging method based on compressed sensing algorithm, characterized in that, Includes the following steps: The original radar echo signal is divided into sub-apertures to obtain multi-channel original echo signals; The multi-channel raw echo signal is preprocessed in the range Doppler domain, including: multi-channel registration, clutter suppression and radial velocity estimation; an imaging model is constructed on the moving target echo signal obtained after clutter suppression, the moving target imaging model includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated; A method based on Legendre polynomial fitting was used to correct translational motion errors. Sparse reconstruction is performed using the alternating direction multiplier method to obtain the moving target imaging results in the range-Doppler domain. The two-dimensional spatially variable error matrix is ​​introduced into the alternating direction multiplier method algorithm framework to compensate for the two-dimensional spatially variable phase error. Based on the relationship between the range-Doppler domain and the northeast coordinate system, a two-dimensional interpolation method is used to project the moving target imaging results in the range-Doppler domain to obtain the moving target focused imaging results in the ground-range plane.

2. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 1, characterized in that, The moving target imaging model is constructed according to the following steps: Range pulse compression and range-directed fast Fourier transform are performed on the multi-channel raw echo signals to obtain the target's echo fundamental frequency signal: in, , and These represent the bandwidth, carrier frequency, and speed of light of the echo signal, respectively. It represents fast time. Represents slow time; Represents the carrier aircraft to the The slant distance of each scattering center This indicates the total number of scattering points contained in the moving target; j The imaginary unit; The target motion is transformed into an ideal turntable model using the following formula: in, In order to be in The instantaneous change in slant distance caused by rotation at any given moment; and Let x and y be the x and y coordinates of the p-th scattering center; The distance between the reference point and the aircraft; The rotation angle of the carrier aircraft around the ideal circular arc with the target center as the origin on the LOS plane; The rotation angle of the moving target around the origin; The carrier aircraft arrived at the Slope distance of each scattering center Represented as: in, and These represent the radial velocity of the target and the radial velocity of the carrier aircraft, respectively. Indicates translational motion; Will According to Taylor expansion, we get: in, , These represent the first and second spatial variations of distance and azimuth, respectively. Order coefficient; Represents any time Projection onto the reference imaging projection plane; After Taylor was unfolded The moving target imaging model is obtained by incorporating the target's echo fundamental frequency signal: 。 3. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 2, characterized in that, The multi-channel raw echo signal is preprocessed in the range Doppler domain, including the following steps: Azimuth compression is achieved by performing a fast Fourier transform on the multi-channel raw echo signal in the azimuth dimension to obtain the range Doppler echo; A multi-channel moving target detection algorithm is used to process the range Doppler echo, and clutter-suppressed moving target echo signal and target radial velocity estimation results are obtained.

4. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 2, characterized in that, The method for correcting translational motion errors based on Legendre polynomial fitting includes the following steps: Extracting the moving target echo phase from the moving target imaging model: in, Represents phase extraction; Represents phase along Untangling; This indicates a fast Fourier transform along the fast time axis; Phase error fitting using Legendre polynomials: in, Represents the number of samples along the distance direction; Representing Legendre polynomials, Represents the coefficients of a polynomial; This represents the estimated residual translational error; For distance to the first Each sampling frequency point The sampling frequency is the range direction. Construct the compensation function: ； The translational motion error is compensated using the following formula, resulting in the moving target imaging model after translational motion error compensation: in, This indicates the inverse fast Fourier transform along the fast time frequency.

5. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 4, characterized in that, The method of using alternating direction multipliers for sparse reconstruction to obtain the moving target imaging results in the range-Doppler domain includes the following steps: The moving target imaging model after translational motion error compensation is divided into blocks along the azimuth dimension in the range-Doppler domain: in, Represents pulse duration; Represents convolution operation; Represents the Doppler frequency; This represents the phase error in the range-Doppler domain; It is a window function along the orientation dimension. Represents the azimuth sampling frequency. Represents the number of blocks. Represents the Doppler frequency interval; Represents the number of blocks; Represents the number of sampling points in the azimuth direction; right Inverse Fast Fourier Transform along the azimuth dimension yields the block data in the range-azimuth-time domain. ; Using the alternating direction multiplier method, for The solution is performed to obtain well-focused segmented sub-images; The moving target imaging results in the range-Doppler domain are obtained by summing up all the segmented sub-images.

6. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 5, characterized in that, The method employing alternating direction multipliers is used for... The solution process yields well-focused sub-image blocks, including the following steps: The solution process is broken down into the following three parts, where the first part is... m The first position sub-block k +1 iterations can be represented as: in, express The target image can be represented in matrix form as... ; Represents the Fourier matrix along the azimuth direction; Represents conjugate transpose; Represents auxiliary variables; Represents the dual variable; Represents the regularization coefficient; Represents a threshold function; Represents the spatially variable phase compensation matrix; In the formula Represented as: in, Representing the Each distance unit, with phase compensation factor along the azimuth dimension; The phase compensation factor for each distance cell is solved using the following formula: in, Representing the n An image of a distance unit; Representing the image number n The distance unit below the first i Values ​​of each azimuth unit; Indicates the first n Echoes from each distance cell, Indicates the first i The Fourier matrix corresponding to each square Doppler frequency value; The phase compensation factor is processed using the image sharpest criterion: in, Indicates the sharpness of an image; The phase compensation factor for each distance cell is solved using the image sharpest criterion. By fitting the data along the range direction, the estimated phase error values ​​corresponding to all range cells are obtained. ; Will Incorporating the alternating direction multiplier algorithm framework In the process of correcting the two-dimensional spatial phase error, after multiple iterations using the alternating direction multiplier method, a well-focused segmented sub-image is obtained.

7. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 5, characterized in that, The imaging results of the moving target in the range-Doppler domain are expressed as follows: 。 8. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 7, characterized in that, The method of projecting the moving target imaging results in the range-Doppler domain onto the ground-based coordinate system using two-dimensional interpolation based on the relationship between the range-Doppler domain and the northeast coordinate system to obtain the moving target focused imaging results in the ground-based distance plane includes the following steps: Based on the target radial velocity estimate Constructing Doppler offset values: ； Translate the image along the Doppler curve: ； Based on the relationship between the distance Doppler domain and the northeast coordinate system, two-dimensional interpolation is used to... Projected to ,in, x The northward distance is obtained by projection. y The eastward distance obtained by projection yields the focused imaging result of the moving target on the ground distance plane.

9. The multi-angle SAR moving target imaging method based on compressed sensing algorithm as described in claim 8, characterized in that, The relationship between the distance-Doppler domain and the northeast coordinate system is as follows: in, , , , These are the aircraft's location and speed information recorded by GPS.

10. A multi-angle SAR moving target imaging device based on compressed sensing algorithm, characterized in that, include: The first module is used to divide the original radar echo signal into sub-apertures to obtain multi-channel original echo signals. The second module is used to preprocess the multi-channel raw echo signal in the range Doppler domain, including: multi-channel registration, clutter suppression and radial velocity estimation; and to construct an imaging model for the moving target echo signal obtained after clutter suppression. The moving target imaging model includes non-vacuum-variable translational motion and two-dimensional vacuum-variable rotational motion that need to be compensated. The third module is used to correct translational motion errors using a method based on Legendre polynomial fitting. The fourth module is used to perform sparse reconstruction using the alternating direction multiplier method to obtain the moving target imaging results in the range-Doppler domain. It introduces the two-dimensional spatially variable error matrix into the alternating direction multiplier method algorithm framework to compensate for the two-dimensional spatially variable phase error. The fifth module is used to project the moving target imaging results in the range-Doppler domain using two-dimensional interpolation based on the relationship between the range-Doppler domain and the northeast coordinate system, so as to obtain the moving target focused imaging results in the ground-range plane.

Images