Chromatographic SAR imaging method combining homonymy point focusing and subspace spectrum estimation

Abstract

The invention discloses a tomography SAR imaging method combining homonymy point focusing and subspace spectrum estimation. The tomography SAR imaging method comprises the following steps: calculating weights and scores based on Euclidean distances, complex coherence coefficients and phase correlation between a main passing image and a plurality of other passing images; determining a registered data matrix according to the weight calculation score, and performing sliding window covariance matrix smoothing processing according to the registered data matrix to obtain a smoothed covariance matrix; performing symmetric processing on the smoothed covariance matrix to construct a weighting factor; performing eigenvalue decomposition on the smoothed covariance matrix to construct a noise subspace; based on the measurement matrix, the weighting factor and the noise subspace, spectrum estimation is carried out, and tomography SAR imaging which keeps good homonymy point focusing precision and three-dimensional imaging quality is completed.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of radar, and particularly relates to a tomographic SAR (Synthetic Aperture Radar) imaging method combining the same-name point focusing and subspace spectrum estimation. BACKGROUND

[0002] With the acceleration of urbanization, the synthetic aperture radar plays an increasingly important role in high-resolution remote sensing monitoring, especially in the environment with complex urban building structures and variable ground object scattering characteristics. The three-dimensional tomographic SAR (TomoSAR) technology has become an important means for identifying and reconstructing targets.

[0003] In the traditional tomographic synthetic aperture radar method, the observation dimension is mainly established in the vertical direction, i.e. the elevation direction, and spectrum estimation or compressive sensing technology is used to reconstruct the target scene in three dimensions. Common spectrum estimation methods include non-parametric methods based on fast Fourier transform (FFT), beam forming (BF), minimum variance (Capon), and multiple signal classification (MUSIC, Multiple Signal Classification).

[0004] Tomographic SAR imaging is based on two-dimensional SAR observation data under multiple orbits and different viewing angles to realize the spatial reconstruction of scatterers in the elevation direction. Its imaging performance largely depends on the high-precision registration between multi-pass SAR images. In practical applications, there are registration errors between airborne multi-pass SAR data, especially in areas with dramatic changes in viewing angles and complex scattering characteristics. Simply relying on image registration cannot achieve pixel-level alignment, which seriously affects the quality of three-dimensional imaging. Therefore, how to introduce a more robust same-name point extraction mechanism and effective noise suppression strategy in the tomographic inversion process has become a key technical challenge to improve the reconstruction accuracy of TomoSAR.

[0005] Among them, the multiple signal classification algorithm is a high-resolution spectrum estimation method based on feature subspace decomposition. Because it has super-resolution capability and can effectively separate coherent scatterers, it is a more refined tomographic focusing technology in complex urban environments. The MUSIC method uses the covariance matrix of the received signal for eigenvalue decomposition, separates the signal subspace and noise subspace orthogonally, and constructs a response vector in the elevation search space. Through spectrum function search to identify the position of the scatterer, it has higher elevation resolution and noise resistance.

[0006] The traditional MUSIC tomography method mainly includes: taking an image of a voyage as a main image, using image registration technology to accurately register the images of other voyages to the reference coordinate system of the main image, and ensuring the consistency of the same scattering point in the image domain under different viewing angles. For any pixel point (x, y) in the main image, the corresponding slant range difference model and elevation response vector are constructed in combination with the track parameters, baseline configuration and wavelength of different voyages. The signal matrix is constructed from the registered multi-voyage complex pixel values, and the covariance matrix is calculated. The spectral function shows a sharp peak at the true scattering body elevation. The elevation positions of multiple main scattering bodies are estimated through spectral peak detection, and super-resolution three-dimensional imaging is realized. The above steps are repeated for all pixel points in the image to obtain the MUSIC elevation inversion result pixel by pixel, and finally a three-dimensional point cloud image is reconstructed.

[0007] The traditional MUSIC algorithm usually relies on the original covariance matrix for eigenvalue decomposition, is easily affected by noise, data disturbance and covariance matrix estimation error, leads to unstable spectral estimation, especially in the case of low signal-to-noise ratio or limited sampling data. In addition, the traditional MUSIC method lacks sufficient use of data structure and prior information, and does not use additional smoothing or enhancement mechanism to improve the quality of the covariance matrix, thereby affecting the resolution of weak scatterers or close-range targets. At the same time, in the context of tomographic SAR imaging with viewing angle changes and speckle noise, direct use of traditional MUSIC can easily cause positioning errors and false peaks, affecting the final imaging accuracy. Secondly, the focusing accuracy of the same named point is often unstable, which seriously affects the imaging quality.

[0008] Therefore, how to provide a tomographic SAR imaging method that maintains good focusing accuracy of the same named point and three-dimensional imaging quality has become an important problem. SUMMARY

[0009] In order to solve the above problems existing in the prior art, the present application provides a tomographic SAR imaging method combining same named point focusing and subspace spectral estimation.

[0010] The technical problem to be solved by the present application is solved by the following technical scheme:

[0011] In a first aspect, the present application provides a tomographic SAR imaging method combining same named point focusing and subspace spectral estimation, which comprises:

[0012] Calculating a weight score based on the Euclidean distance, complex coherence coefficient and phase correlation between the main image and the other images.

[0013] According to the weight score, a registered data matrix is determined, and a sliding window covariance matrix smoothing process is performed on the registered data matrix to obtain a smoothed covariance matrix.

[0014] Performing symmetric processing on the smoothed covariance matrix to construct weighting factors;

[0015] Performing eigenvalue decomposition on the smoothed covariance matrix to construct a noise subspace;

[0016] Spectrum estimation is performed based on the measurement matrix, the weighting factors and the noise subspace to complete three-dimensional tomographic SAR imaging.

[0017] Optionally, a score is calculated based on the Euclidean distance, complex coherence coefficient, and phase correlation between the primary flyby image and multiple other flyby images, including:

[0018] weighting the Euclidean distance, complex coherence coefficient, and phase correlation between the primary flyby image and the plurality of other flyby images according to an initial threshold and preset weights to obtain an intermediate weighted calculation score; the preset weights include a preset weight for the Euclidean distance, a preset weight for the complex coherence coefficient, and a preset weight for the phase correlation;

[0019] The step of iteratively updating and determining the intermediate weight calculation score, until the intermediate weight calculation score is less than a preset threshold, determines the intermediate weight calculation score in the current iteration as the weight calculation score.

[0020] Optionally, the preset weight is determined by:

[0021]

[0022] Wherein, i∈{1,2,3}, λ1 represents the preset weight of the Euclidean distance, λ2 represents the preset weight of the complex coherence coefficient, and λ3 represents the preset weight of the phase correlation; represents the λ of the tth iteration i ; represents the λ of the t+1th iteration i ; ΔS represents the gradient of the intermediate weight calculation score; η represents the learning rate.

[0023] Optionally, performing symmetric processing on the smoothed covariance matrix to construct a weighting factor includes:

[0024] Performing symmetric processing on the antisymmetric permutation matrix and the smoothed covariance matrix to obtain a symmetric covariance matrix;

[0025] The symmetric covariance matrix is ​​processed by processing a diagonal matrix function to construct a weighting factor.

[0026] Optionally, performing spectrum estimation based on the measurement matrix, the weighting factor, and the noise subspace to complete three-dimensional tomographic SAR imaging includes:

[0027] The measurement matrix, the weighting factor and the noise subspace are substituted into a MUSIC spectrum calculation formula to perform spectrum estimation, and three-dimensional tomographic SAR imaging is completed.

[0028] Optionally, the MUSIC spectrum calculation formula comprises:

[0029]

[0030] wherein P MUSIC (θ i ) represents a spectrum estimation value of an i th elevation candidate point θ i ; a(θ i ) represents an i th column in the measurement matrix, representing a direction vector of the i th elevation candidate point θ i ; W represents the weighting factor; E n represents the noise subspace; and a superscript H represents a conjugate transpose operation of a matrix.

[0031] In a second aspect, the present application provides a tomographic SAR imaging device combining the same-named point focusing and subspace spectrum estimation, comprising:

[0032] A calculation module is configured to calculate a weight calculation score based on a Euclidean distance, a complex coherence coefficient and a phase correlation between a main stripmap image and multiple other stripmap images.

[0033] A smoothing module is configured to determine a registered data matrix according to the weight calculation score, and perform sliding window covariance matrix smoothing processing according to the registered data matrix to obtain a smoothed covariance matrix.

[0034] A symmetry processing module is configured to perform symmetry processing on the smoothed covariance matrix to construct a weighting factor.

[0035] An eigenvalue decomposition module is configured to perform eigenvalue decomposition on the smoothed covariance matrix to construct a noise subspace.

[0036] A spectrum estimation module is configured to perform spectrum estimation based on a measurement matrix, the weighting factor and the noise subspace to complete three-dimensional tomographic SAR imaging.

[0037] The application provides a tomographic SAR imaging method combining same-name point focusing and subspace spectrum estimation.

[0038] The application will be further described in detail below with reference to the drawings and the application. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 FIG. 1 is a flowchart of a tomographic SAR imaging method combining same-name point focusing and subspace spectrum estimation according to an embodiment of the application;

[0040] Figure 2 FIG. 2 is a flowchart of adaptive same-name point focusing according to an embodiment of the application;

[0041] Figure 3 FIG. 3 is a flowchart of subspace reinforcement according to an embodiment of the application;

[0042] Figure 4 FIG. 4 is a first imaging comparison diagram of a tomographic SAR imaging method and an existing tomographic SAR imaging method according to an embodiment of the application;

[0043] Figure 5 FIG. 5 is a second imaging comparison diagram of a tomographic SAR imaging method and an existing tomographic SAR imaging method according to an embodiment of the application;

[0044] Figure 6 FIG. 6 is a structural diagram of a tomographic SAR imaging device combining same-name point focusing and subspace spectrum estimation according to an embodiment of the application. DETAILED DESCRIPTION

[0045] The application will be further described in detail below with reference to the drawings and the application.

[0046] In order to solve the problem that the focusing accuracy of the same name point is often unstable and the imaging accuracy is poor in the existing tomographic SAR imaging method, an embodiment of the present application provides a tomographic SAR imaging method combining the focusing of the same name point and the subspace spectrum estimation, as shown in Figure 1 , Figure 1 is a flowchart of the tomographic SAR imaging method combining the focusing of the same name point and the subspace spectrum estimation provided by the embodiment of the present application, and specifically includes the following steps:

[0047] In step S101, a score is calculated based on the Euclidean distance, the complex coherence coefficient and the phase correlation between the main stripmap image and other stripmap images.

[0048] In the embodiment of the present application, stripmap refers to the process of continuously imaging a certain area on the ground by a SAR platform such as a satellite or an airplane flying along a predetermined orbit. Stripmap image refers to a strip-shaped or special-shaped radar image generated by each stripmap. The main stripmap image refers to a selected stripmap image, and the other stripmap images refer to the other stripmap images except the main stripmap image.

[0049] Referring to Figure 2 , Figure 2 is a flowchart of the adaptive focusing of the same name point provided by the embodiment of the present application, and a score is calculated based on the Euclidean distance, the complex coherence coefficient and the phase correlation between the main stripmap image and multiple other stripmap images, including:

[0050] The Euclidean distance, the complex coherence coefficient and the phase correlation between the main stripmap image and multiple other stripmap images are weighted and distributed according to the initial threshold and the preset weight to obtain an intermediate weight calculation score. The preset weight includes the preset weight of the Euclidean distance, the preset weight of the complex coherence coefficient and the preset weight of the phase correlation.

[0051] The step of iteratively updating and determining the intermediate weight calculation score is performed until the intermediate weight calculation score is less than the preset threshold, and the intermediate weight calculation score under the current iteration is determined as the weight calculation score.

[0052] The determination method of the preset weight includes:

[0053] First, the initial weight is set, and the iterative weight updating method is:

[0054]

[0055] wherein i is an element of {1, 2, 3}, λ1 represents the preset weight of the Euclidean distance, λ2 represents the preset weight of the complex coherence coefficient, and λ3 represents the preset weight of the phase correlation; the superscript t represents the number of iterations, represents λ i of the t+1th iteration. λ represents the t-th iteration i ; ΔS represents an intermediate weight calculation score, i.e., a gradient of the same point focusing score; η represents a learning rate (decreased with iterations).

[0056] The weight updating strategy provided by the embodiment of the application is simple in form, only involves multiplication and addition operations, is convenient to calculate, and is suitable for efficient implementation in large-scale data processing tasks. In mechanism, the method can dynamically adjust the weight according to the change of the score function, avoid the rigid problem caused by fixed weight updating, and has good flexibility and adaptability. With the iteration, the algorithm can gradually enhance the response ability to high-contribution features, effectively suppress invalid or noise features, and thus improve the same point focusing accuracy and robustness. In addition, the weight evolution process reflects the change trend of the data structure with iterations, which helps the algorithm to realize stable convergence in the search space and improve the overall imaging quality and algorithm reliability.

[0057] When the number of iterations reaches a preset number, the weight under the iteration is taken as a preset weight. The preset number can be set by the technical personnel according to experience.

[0058] In the embodiment of the application, a preset threshold is determined. In SAR tomographic imaging, a single same point focusing index is difficult to comprehensively measure the same point focusing quality. Therefore, an overall score function is used to comprehensively consider the spatial geometric consistency, complex coherence coefficient and coherent scattering characteristics, so as to ensure that the same point reaches a global optimum.

[0059] The manner of calculating the weight calculation score is as follows:

[0060] score = λ1 x Euc + λ2 x coh + λ3 x pha;

[0061] Wherein, score represents an intermediate weight calculation score; λ1 represents a preset weight of the Euclidean distance; Euc represents the Euclidean distance; λ2 represents a preset weight of the complex coherence coefficient; coh represents the complex coherence coefficient; λ3 represents a preset weight of the phase correlation; and pha represents the phase correlation.

[0062] The minimum search is used:

[0063] best match = argmin score;

[0064] Wherein, best match represents a best matching operation; and argmin is used to represent a parameter value of a function that obtains a minimum value in its definition domain.

[0065] When the intermediate weight calculation score is less than a preset threshold, i.e., lower than a preset standard, the iteration is terminated, the intermediate weight calculation score under the current iteration is determined as the optimal weight calculation score, and the pixel point in the main overflight image at this time is the homonym of the pixel in the image of other overflights.

[0066] The traditional algorithm usually only focuses on the registration process of the image, ignores the subtle differences of different overflight data in scattering characteristics, view angle changes and the like, and is prone to cause mismatching in local areas. In contrast, the iterative optimization mechanism is introduced in the embodiment of the application, the matching adjustment strategy at the pixel level is combined, the focusing weight of the homonym is dynamically updated, the matching process gradually converges to the optimal solution under multi-scale and multi-feature constraints, and therefore the accuracy and robustness of the homonym focusing are significantly improved. The dynamic weighting homonym focusing framework based on the iterative optimization mechanism is designed in the application, the weight of each feature factor is updated round by round, and the homonym focusing result gradually converges to the global optimal solution. The method is particularly suitable for complex SAR imaging scenes such as dense buildings and forest shelter, and effectively improves the robustness of the algorithm to non-ideal conditions. The iteration is automatically terminated when the error is lower than the preset standard, redundant calculation is effectively avoided, and the overall operation efficiency is significantly improved.

[0067] In the embodiment of the application, in the iterative optimization process, the Euclidean distance is used as a key constraint condition to ensure the spatial geometric consistency of the homonym, effectively avoid geometric mismatching, and improve the spatial constraint accuracy.

[0068] Geometric constraint assisted homonym focusing: the Euclidean distance can constrain the three-dimensional position relationship of the homonym and suppress the deviation caused by noise or local disturbance.

[0069] The calculation formula of the Euclidean distance is as follows:

[0070]

[0071] Wherein, amp n (r, x) is a pixel point with a position of (r, x) in the selected main overflight image; amp uv (r, x) is a pixel point with a position of (r, x) in the other overflight image; r is the horizontal coordinate of the pixel point in the overflight image, x is the vertical coordinate of the pixel point in the overflight image; N represents the total number of overflights; and L1 represents the size of the overflight image.

[0072] In the embodiment of the application, in the SAR tomographic imaging, the homonym not only needs to be aligned at the pixel level, but also needs to be correct in the physical space. The complex coherence coefficient, as an important index for measuring the similarity of the homonym, can effectively improve the rationality of the focusing of the target homonym.

[0073] The complex coherence coefficient is used to measure the consistency of scatterer responses in SAR observations from different flights. High coherence means stable scattering characteristics, which helps improve the credibility of homonymous points.

[0074] The calculation formula of the complex coherence coefficient is:

[0075]

[0076] Where Coh represents the complex coherence coefficient.

[0077] In an embodiment of the present invention, in SAR tomography, the scattering characteristics of the same-name points are mainly reflected in the phase information. Therefore, introducing phase correlation as a same-name point focusing constraint helps to improve the focusing degree of the coherent information and make the same-name point focusing process more accurate and robust.

[0078] In SAR images from different flights, the amplitude may vary due to changes in imaging conditions, but phase information is more stable. Leveraging phase correlation for focusing on homonymous points ensures that the scattering characteristics of these homonymous points remain physically consistent.

[0079] Compared with directly using amplitude homonymous point focusing, phase homonymous point focusing is more resistant to coherent speckle noise interference, improves the robustness of homonymous point focusing, and enhances noise resistance.

[0080] Compared with traditional methods, this strategy performs better in the task of focusing on the same-name points of complex or non-rigid targets such as urban buildings and forests. It can adapt to the scattering characteristics of different targets, is suitable for complex scenes, and can improve the accuracy of three-dimensional reconstruction.

[0081] The calculation formula for phase correlation is:

[0082]

[0083] (u=1,2,…,N,v=1,2,…,L1,u≠n)

[0084] Where Pha represents the phase correlation; R n (m) is the reference slant distance of the selected main flyby image, R p (m) is the reference slant distance of other flyby images.

[0085] In the embodiment of the present invention, Euclidean distance is used to constrain spatial geometric positions, prevent geometric mismatch, and integrate multi-pixel information to ensure accurate spatial alignment of homonymous points;

[0086] By constraining the complex coherence coefficient, we ensure that the same-name points meet the scattering consistency, improve the accuracy, and enhance the precision of tomographic imaging;

[0087] The scattering characteristic consistency of a target in multi-view imaging is utilized by using phase correlation measurement, phase homonymy point focusing is adopted, coherent information is utilized to improve the robustness of homonymy point focusing, reduce the influence of noise, and enhance the stability of homonymy points between cross-sequence images.

[0088] In step S102, the registered data matrix is determined according to the score calculated by the weight, and the sliding window covariance matrix smoothing processing is performed according to the registered data matrix, to obtain the smoothed covariance matrix.

[0089] Referring to Figure 3 , Figure 3 is a flowchart of the subspace enhancement provided by the embodiment of the application, the homonymy points of the flight are obtained according to the score calculated by the weight, the complex values of the homonymy points of the flight are stored, and the registered data matrix is obtained wherein M indicates the sampling point number in the elevation direction, each row in the registered data matrix is a channel / subaperture, and each column is a fast time or frequency point.

[0090] Supposing that the sliding window length is L=N, K=N-L+1 times of sliding are performed in total, and the kth subarray is:

[0091]

[0092] The corresponding covariance matrix is:

[0093]

[0094] Finally, the smoothed covariance matrix after the sliding window covariance matrix smoothing processing is:

[0095]

[0096] The sliding subarray covariance matrix smoothing processing effectively improves the statistical stability of the covariance matrix by weighted average of multiple local subarrays, and significantly suppresses the interference caused by high-frequency noise and local mutation. The method reduces the influence of random noise on spectrum estimation through fusion of local redundant information, thereby enhancing the robustness and resolution capability of spectrum peak identification.

[0097] In step S103, the symmetric processing is performed on the smoothed covariance matrix to construct a weighting factor.

[0098] The traditional method has no symmetry, and there may be numerical errors of the non-Hermitian matrix of the covariance matrix. In the embodiment of the application, the forward and backward smoothing fusion can be realized by introducing the anti-symmetric permutation matrix, the Hermitian property and the numerical stability of the covariance matrix are effectively improved, not only the orthogonality of the signal and noise subspaces is enhanced, but also the interference of non-ideal items on the imaging quality is significantly suppressed, which is particularly suitable for accurate three-dimensional reconstruction in low signal-to-noise ratio and complex scattering environment.

[0099] In an implementation, the symmetric processing is performed on the smoothed covariance matrix to construct the weighting factor, comprising:

[0100] The symmetric processing is performed on the anti-symmetric permutation matrix and the smoothed covariance matrix to obtain a symmetrized covariance matrix;

[0101] The symmetrized covariance matrix is processed by processing a diagonal matrix function to construct the weighting factor.

[0102] Specifically, the aforementioned R smooth is a smoothed covariance matrix, and the dimension is N*N. The size of the anti-symmetric permutation matrix J is based on R smooth is determined, and the anti-symmetric permutation matrix J is:

[0103]

[0104] After the smoothed covariance matrix and the anti-symmetric permutation matrix are obtained, the symmetrized covariance matrix R1 is determined in the following manner:

[0105]

[0106] On this basis, the symmetrized covariance matrix is processed by processing a diagonal matrix function diag(·) to construct a weighting factor W, comprising:

[0107]

[0108] wherein, represents a real number set.

[0109] In the embodiment of the application, the weighting factor can suppress high-energy noise channels, improve main lobe focusing and side lobe suppression capability, and is particularly effective in a sparse / occluded structure environment.

[0110] The Hermitian symmetrization processing is introduced to construct the inverse covariance matrix, which effectively strengthens the positive definiteness and symmetry of the matrix, and helps to construct a purer noise subspace and improve the spectral peak resolution capability in a multi-target scene.

[0111] In step S104, eigenvalue decomposition is performed on the smoothed covariance matrix to construct a noise subspace.

[0112] In the embodiment of the application, the eigenvalue decomposition performed on the smoothed covariance matrix comprises:

[0113]

[0114] wherein, U s represents an eigenvector matrix of a signal subspace; Λ s represents an eigenvalue matrix corresponding to the signal subspace; U Nthe eigenvector matrix representing the noise subspace; Λ N the eigenvalue matrix corresponding to the noise subspace; the superscript H represents the conjugate transpose operation of the matrix.

[0115] Based on the above eigenvalue decomposition result, the signal space and the noise space are separated:

[0116] Suppose that there are d signal sources in the observation signal, then the first d largest eigenvalues correspond to the signal subspace, and the remaining M-d smaller eigenvalues correspond to the noise subspace E n :

[0117] E n = [e d+1 ,e d+2 ,…e M ];

[0118] By extracting the principal components of the signal, the signal and the noise can be accurately distinguished. The extraction of the noise subspace En can significantly improve the spectrum estimation result, so that the MUSIC method is more accurate. The role of the noise subspace is to reduce the influence of noise, so as to more accurately identify the scatterers.

[0119] In step S105, spectrum estimation is performed based on the measurement matrix, the weighting factor and the noise subspace, and three-dimensional tomographic SAR imaging is completed.

[0120] The measurement matrix is mainly used to obtain low-dimensional observation data from high-dimensional observation signals, and is a key bridge for mapping the distribution of scatterers in the height direction to echo data under multiple observation angles.

[0121] In the embodiment of the application, the measurement matrix, the weighting factor and the noise subspace are substituted into the MUSIC spectrum calculation formula for spectrum estimation, which includes:

[0122]

[0123] where P MUSIC (θ i ) represents the spectrum estimation value of the i-th column in the measurement matrix; a(θ i ) represents the i-th column in the measurement matrix, representing the direction vector of the i-th direction θ i or the elevation candidate point θ i ; the superscript H represents the conjugate transpose operation of the matrix.

[0124] In tomographic SAR, the three-dimensional structure of the target can be estimated according to the spectrum value, and tomographic SAR imaging is realized.

[0125] In the embodiment of the present application, the score is calculated based on the Euclidean distance, complex coherence coefficient and phase correlation between the main flight image and the plurality of other flight images, the high-precision common point focusing is realized by comprehensively considering the Euclidean distance, phase correlation and complex coherence coefficient characteristics, the statistical stability of the covariance matrix is effectively improved by using the sliding window covariance matrix smoothing processing, and the interference caused by high-frequency noise and local mutation is significantly suppressed, the robustness under the condition of signal-to-noise ratio is improved by introducing the weighted factor, the main lobe strength is improved, and the side lobe is suppressed. Based on this, the tomographic SAR imaging method provided by the embodiment of the present application combines the common point focusing and subspace spectrum estimation, and can still maintain good common point focusing precision and three-dimensional imaging quality when facing complex scenes such as cities, thereby providing theoretical support and engineering feasibility basis for high-precision TomoSAR imaging tasks.

[0126] The simulation experiment of the tomographic SAR imaging method combining common point focusing and subspace spectrum estimation provided by the embodiment of the present application is as follows:

[0127] Referring to Figure 4 and Figure 5 , Figure 4 is the first imaging comparison schematic diagram of the tomographic SAR imaging method provided by the embodiment of the present application and the existing tomographic SAR imaging method, Figure 5 is the second imaging comparison schematic diagram of the tomographic SAR imaging method provided by the embodiment of the present application and the existing tomographic SAR imaging method, wherein the existing tomographic SAR imaging method includes MUSIC, Capon, CS-OMP, ANM and ISTA. Figure 4 and Figure 5 It can be known that the imaging effect and precision of the tomographic SAR imaging method provided by the embodiment of the present application are obviously better than those of the existing tomographic SAR imaging method.

[0128] Referring to Table 1, Table 1 is the index comparison table of the tomographic SAR imaging method provided by the embodiment of the present application and the existing tomographic SAR imaging method, and the details are as follows:

[0129] Table 1 index comparison table

[0130] Discrepancy K-neighbor point cloud entropy The method 0.0412 3.7049 MUSIC 0.1248 4.0158 Capon 0.0827 4.0441 CS-OMP 0.0779 3.8654 ANM 0.1209 3.8889 ISTA 0.0854 3.7434

[0131] The present application proposes an adaptive weighting mechanism, which effectively enhances the main lobe energy focusing, suppresses the side lobe leakage and pseudo-peak interference according to the local signal-to-noise characteristic distribution in the spectrum estimation process, so as to improve the resolution capability and imaging stability of the overall spectrum diagram.

[0132] The application introduces a same-name point focusing scoring mechanism of fusion features, on the basis of a traditional image registration strategy, full-dimensional constraints are performed through construction of a comprehensive scoring function (including Euclidean distance, complex coherence coefficient, phase correlation), same-name point focusing error is significantly reduced, accurate positioning of the same-name points is realized, then covariance matrix smoothing processing is performed through a sliding subarray, a purer noise subspace is constructed, adaptive weighting is performed for multiple signal classification three-dimensional tomographic SAR imaging, finally, point cloud reconstruction quality is significantly improved, and height estimation error is effectively reduced. Among them, the dispersion reaches 0.0412, which is reduced by nearly half compared with the optimal result 0.0779 of other methods, and higher focusing accuracy is embodied; the K-neighbor point cloud entropy is 3.7049, which is better than the minimum value 3.8654 of other methods, and the advantages of the method in structural clarity and information integration are fully proved.

[0133] Based on the same inventive concept, the application embodiment further provides a tomographic SAR imaging device combining same-name point focusing and subspace spectrum estimation, referring to Figure 6 , Figure 6 is a structural schematic diagram of a tomographic SAR imaging device combining same-name point focusing and subspace spectrum estimation provided by the application embodiment, and the tomographic SAR imaging device comprises:

[0134] The computing module 601 is configured to calculate a weight calculation score based on Euclidean distance, complex coherence coefficient and phase correlation between a main flight image and multiple other flight images.

[0135] The smoothing module 602 is configured to determine a registered data matrix according to the weight calculation score, and perform sliding window covariance matrix smoothing processing according to the registered data matrix to obtain a smoothed covariance matrix.

[0136] The symmetry processing module 603 is configured to perform symmetry processing on the smoothed covariance matrix to construct a weighting factor.

[0137] The eigenvalue decomposition module 604 is configured to perform eigenvalue decomposition on the smoothed covariance matrix to construct a noise subspace.

[0138] The spectrum estimation module 605 is configured to perform spectrum estimation based on a measurement matrix, the weighting factor and the noise subspace to complete three-dimensional tomographic SAR imaging.

[0139] In an embodiment of the present invention, a score is calculated based on the weights of the Euclidean distance, complex coherence coefficient, and phase correlation between the main flyby image and multiple other flyby images. By combining the characteristics of the Euclidean distance, phase correlation, and complex coherence coefficient, high-precision same-name point focusing is achieved. By adopting a sliding window covariance matrix smoothing method, the statistical stability of the covariance matrix is ​​effectively improved, and the interference caused by high-frequency noise and local mutations is significantly suppressed. By introducing weighting factors, the robustness under the signal-to-noise ratio can be improved, the main lobe intensity can be increased, and the side lobes can be suppressed. Based on this, the tomographic SAR imaging method of combined same-name point focusing and subspace spectrum estimation provided by the embodiment of the present invention can still maintain good same-name point focusing accuracy and three-dimensional imaging quality when facing complex structural scenes such as cities, providing theoretical support and engineering feasibility basis for high-precision TomoSAR imaging tasks.

[0140] Optional computing module, specifically used for:

[0141] According to the initial threshold and preset weights, the Euclidean distance, complex coherence coefficient and phase correlation between the main flyby image and multiple other flyby images are weighted to obtain an intermediate weight calculation score; the preset weights include the preset weights of the Euclidean distance, the preset weights of the complex coherence coefficient and the preset weights of the phase correlation; the step of iteratively updating the intermediate weight calculation score is performed until the intermediate weight calculation score is less than the preset threshold, and the intermediate weight calculation score under the current iteration is determined as the weight calculation score.

[0142] Optionally, the preset weight is determined by:

[0143]

[0144] Wherein, i∈{1,2,3}, λ1 represents the preset weight of the Euclidean distance, λ2 represents the preset weight of the complex coherence coefficient, and λ3 represents the preset weight of the phase correlation; represents the λ of the tth iteration i ; represents the λ of the t+1th iteration i ; ΔS represents the gradient of the intermediate weight calculation score; η represents the learning rate.

[0145] Optional symmetric processing module, specifically used for:

[0146] Symmetric processing is performed on the antisymmetric permutation matrix and the smoothed covariance matrix to obtain a symmetric covariance matrix; the symmetric covariance matrix is ​​processed by processing a diagonal matrix function to construct a weighting factor.

[0147] Optional spectrum estimation module, specifically used for:

[0148] The measurement matrix, the weighting factor and the noise subspace are substituted into a MUSIC spectrum calculation formula to perform spectrum estimation, and three-dimensional tomographic SAR imaging is completed.

[0149] Optionally, the MUSIC spectrum calculation formula comprises:

[0150]

[0151] wherein P MUSIC (θ i ) represents a spectrum estimation value of an i th elevation candidate point θ i ; a(θ i ) represents an i th column in the measurement matrix, representing a direction vector of the i th elevation candidate point θ i ; W represents the weighting factor; E n represents the noise subspace; and a superscript H represents a conjugate transpose operation of a matrix.

[0152] It should be noted that the terms "first", "second", and so on are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Rather, they are merely examples of devices and methods consistent with some aspects of the present application.

[0153] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like means that the specific features or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in the present specification.

[0154] Although the present application is described herein in conjunction with various embodiments, other variations of the disclosed embodiments can be understood and implemented by those skilled in the art with reference to the drawings and the disclosure. In the description of the present application, the word "comprising" does not exclude other components or steps, "one" or "an" does not exclude a plurality, and "plurality" means two or more, unless otherwise explicitly specified. In addition, some measures are described in different embodiments, but this does not mean that these measures cannot be combined to produce good results.

[0155] For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant part can be seen in the part of the method embodiment.

[0156] It should be noted that the device of the embodiment of the present application is the device applying the above-mentioned tomographic SAR imaging method combining the same point focusing and subspace spectrum estimation, and all the embodiments of the above-mentioned tomographic SAR imaging method combining the same point focusing and subspace spectrum estimation are applicable to the device, and can achieve the same or similar beneficial effects.

[0157] The above is a further detailed description of the present application in combination with specific preferred embodiments, and the specific implementation of the present application cannot be limited to these descriptions. For ordinary skilled persons in the technical field to which the present application belongs, without departing from the concept of the present application, a number of simple deductions or substitutions can be made, which should be considered as falling within the protection scope of the present application.

Claims

1. A tomographic SAR imaging method combining homonymous point focusing and subspace spectrum estimation, characterized in that: The tomographic SAR imaging method comprises: The score is calculated based on the Euclidean distance, complex coherence coefficient and phase correlation between the main flyby image and multiple other flyby images; Determining a registered data matrix according to the weighted calculation score, and performing sliding window covariance matrix smoothing processing on the registered data matrix to obtain a smoothed covariance matrix; Performing symmetric processing on the smoothed covariance matrix to construct weighting factors; Performing eigenvalue decomposition on the smoothed covariance matrix to construct a noise subspace; Spectrum estimation is performed based on the measurement matrix, the weighting factors and the noise subspace to complete three-dimensional tomographic SAR imaging.

2. The tomographic SAR imaging method according to claim 1, wherein: The score is calculated based on the Euclidean distance, complex coherence coefficient, and phase correlation between the main flyby image and multiple other flyby images, including: weighting the Euclidean distance, complex coherence coefficient, and phase correlation between the main flyby image and the plurality of other flyby images according to an initial threshold and preset weights to obtain an intermediate weighted calculation score; the preset weights include a preset weight for the Euclidean distance, a preset weight for the complex coherence coefficient, and a preset weight for the phase correlation; The step of iteratively updating and determining the intermediate weight calculation score, until the intermediate weight calculation score is less than a preset threshold, determines the intermediate weight calculation score in the current iteration as the weight calculation score.

3. The tomographic SAR imaging method according to claim 2, wherein: The method for determining the preset weight includes: Wherein, i∈{1,2,3}, λ1 represents the preset weight of the Euclidean distance, λ2 represents the preset weight of the complex coherence coefficient, and λ3 represents the preset weight of the phase correlation; represents the λ of the t-th iteration i ; represents the λ of the t+1th iteration i ; ΔS represents the gradient of the intermediate weight calculation score; η represents the learning rate.

4. The tomographic SAR imaging method according to claim 1, wherein: The smoothed covariance matrix is ​​symmetrically processed to construct weighting factors, including: Performing symmetric processing on the antisymmetric permutation matrix and the smoothed covariance matrix to obtain a symmetric covariance matrix; The symmetric covariance matrix is ​​processed by processing a diagonal matrix function to construct a weighting factor.

5. The tomographic SAR imaging method according to claim 1, wherein: Performing spectrum estimation based on the measurement matrix, the weighting factor, and the noise subspace to complete three-dimensional tomographic SAR imaging includes: The measurement matrix, the weighting factor and the noise subspace are substituted into the MUSIC spectrum calculation formula to perform spectrum estimation and complete three-dimensional tomographic SAR imaging.

6. The tomographic SAR imaging method according to claim 5, characterized in that: The MUSIC spectrum calculation formula includes: Among them, P MUSIC (θ i ) represents the i-th elevation candidate point θ i Spectral estimate of a(θ i ) represents the i-th column in the measurement matrix, representing the i-th elevation candidate point θ i The direction vector of ; W represents the weighting factor; E n represents the noise subspace; the superscript H represents the conjugate transpose operation of the matrix.

7. A tomographic SAR imaging device combining homonymous point focusing and subspace spectrum estimation, characterized in that: The tomographic SAR imaging device comprises: a calculation module for calculating a score by weighting based on a Euclidean distance, a complex coherence coefficient, and a phase correlation between a main flyby image and a plurality of other flyby images; a smoothing module, configured to determine a registered data matrix according to the weighted calculation score, and perform sliding window covariance matrix smoothing processing on the registered data matrix to obtain a smoothed covariance matrix; A symmetric processing module, configured to perform symmetric processing on the smoothed covariance matrix to construct a weighting factor; an eigenvalue decomposition module, configured to perform eigenvalue decomposition on the smoothed covariance matrix to construct a noise subspace; The spectrum estimation module is used to perform spectrum estimation based on the measurement matrix, the weighting factor and the noise subspace to complete three-dimensional tomographic SAR imaging.

8. The tomographic SAR imaging device according to claim 7, characterized in that: The computing module is specifically configured to: According to the initial threshold and preset weights, the Euclidean distance, complex coherence coefficient and phase correlation between the main flyby image and multiple other flyby images are weighted to obtain an intermediate weight calculation score; the preset weights include the preset weights of the Euclidean distance, the preset weights of the complex coherence coefficient and the preset weights of the phase correlation; the step of iteratively updating and determining the intermediate weight calculation score until the intermediate weight calculation score is less than the preset threshold, determines that the intermediate weight calculation score under the current iteration is the weight calculation score.

9. The tomographic SAR imaging device according to claim 8, characterized in that: The method for determining the preset weight includes: Wherein, i∈{1,2,3}, λ1 represents the preset weight of the Euclidean distance, λ2 represents the preset weight of the complex coherence coefficient, and λ3 represents the preset weight of the phase correlation; represents the λ of the t-th iteration i ; represents the λ of the t+1th iteration i ; ΔS represents the gradient of the intermediate weight calculation score; η represents the learning rate.

10. The tomographic SAR imaging device according to claim 7, characterized in that: The symmetry processing module is specifically used to: Symmetric processing is performed on the antisymmetric permutation matrix and the smoothed covariance matrix to obtain a symmetric covariance matrix; the symmetric covariance matrix is ​​processed by processing a diagonal matrix function to construct a weighting factor.

Images