Sea clutter suppression method and device and storage medium

Abstract

The invention relates to the technical field of radar signal processing, and provides a sea clutter suppression method and device and a storage medium, and the method comprises the steps: obtaining radar echo data, and determining training data which comprises a plurality of training samples; according to the training data, obtaining a sample matrix corresponding to the training sample, the sample matrix being a Toeplitz-Hermite positive definite matrix; according to the sample matrix, based on a Brauer disc theorem, determining a classification boundary; screening the sample matrix according to the classification boundary to obtain a clutter matrix; according to each clutter matrix, on the basis of the constructed subspace manifold, obtaining a clutter covariance matrix; and determining a space-time filter weight vector according to the clutter covariance matrix. A clutter matrix corresponding to a non-target signal is obtained through screening based on the Brauer disk theorem, the self-zero effect is avoided, the sea clutter suppression accuracy is improved, and based on construction of the subspace manifold, structured information of data can be fully utilized, and the sea clutter suppression accuracy is further improved.

Description

Technical Field

[0001] The present invention relates to the field of radar signal processing technology, and in particular to a sea clutter suppression method, device and storage medium. Background Art

[0002] Sea clutter is formed by radar echoes reflected from the sea surface. It has the characteristics of high intensity, wide spectrum, and strong randomness, which will seriously affect the radar's target detection and recognition capabilities. In order to reduce the impact of sea clutter on radar detection results, the existing technology uses the Space-Time Adaptive Processing (STAP) method to suppress the impact of sea clutter. The STAP method constructs a space-time two-dimensional filter by simultaneously considering the spatial and temporal dimensions of the radar echo to achieve the suppression of clutter and noise. The STAP method can adaptively adjust the filter parameters according to different radar environments and mission requirements to achieve the optimal clutter suppression effect. The core problem in the STAP method is the estimation of the clutter covariance matrix, and then obtains the weight vector parameters of the space-time two-dimensional filter.

[0003] However, the existing STAP method is mainly aimed at point targets. For cluster target scenarios, there are extended-range targets, that is, targets occupying multiple range units. The existing STAP method will have a target self-zero effect. The target self-zero effect refers to the mixing of target signals in the training samples, resulting in inaccurate estimation of the clutter covariance matrix, so that the formed space-time two-dimensional filter will suppress the target energy in the unit to be tested, resulting in the loss of output power of the target. In addition, the existing STAP method ignores the geometric properties and structural characteristics of radar data, and cannot use the structural information of the data to estimate the clutter covariance matrix, which affects the accuracy of suppressing sea clutter.

[0004] Therefore, in the case of extended-range targets, how to improve the suppression effect of sea clutter, accurately suppress sea clutter without affecting the target signal, and use the structural information of radar data to improve the accuracy of suppressing sea clutter is an urgent problem to be solved. Summary of the invention

[0005] The present invention provides a sea clutter suppression method, device and storage medium, which are used to solve the problem that in the prior art, the STAP method has the problem of inaccurate estimation of the clutter covariance matrix due to the self-zero effect when there is a range-extended target, and the defect that the structural information of radar data is not used to improve the accuracy of suppressing sea clutter.

[0006] The present invention provides a sea clutter suppression method, comprising: Acquire radar echo data and determine training data, wherein the training data includes a plurality of training samples; According to the training data, a sample matrix corresponding to the training sample is obtained, where the sample matrix is ​​a Toeplitz-Hermitian positive definite matrix; According to the sample matrix, based on the Brauer disk theorem, a classification boundary is determined, wherein the classification boundary represents a boundary between a target signal and a clutter signal; Filter the sample matrix according to the classification boundary to obtain a clutter matrix; According to each of the clutter matrices, based on constructing a subspace manifold, obtaining a clutter covariance matrix; A space-time filter weight vector is determined according to the clutter covariance matrix.

[0007] According to a sea clutter suppression method provided by the present invention, the clutter covariance matrix is ​​obtained based on constructing a subspace manifold according to each of the clutter matrices, including: Decomposing the clutter matrix to obtain a corresponding eigenvector matrix and an eigenvalue diagonal matrix; Constructing a Grassmann manifold according to each of the eigenvector matrices and the corresponding eigenvalue diagonal matrices; The clutter covariance matrix is ​​obtained by performing calculations based on the Grassmann manifold and each of the clutter matrices and based on a volume correlation function.

[0008] According to a sea clutter suppression method provided by the present invention, the step of acquiring radar echo data and determining training data includes: Acquire radar echo data, wherein the radar echo data includes data of array element dimension, pulse dimension and distance dimension; According to the radar echo data, based on the position of the sliding window in the distance dimension, determining a preset number of the training samples to form the training data; The length of the sliding window is a preset value, the sliding window is used to move along the distance dimension and determine the range of data processing according to the position of the sliding window, and the training sample includes a matrix formed by corresponding array element dimension data and pulse dimension data at the distance.

[0009] According to a sea clutter suppression method provided by the present invention, the classification boundary is determined based on the sample matrix and the Brauer disk theorem, including: Performing eigenvalue decomposition on the sample matrix, obtaining corresponding eigenvalue information and determining the maximum Brauer disk center, wherein the maximum Brauer disk center is determined by the maximum eigenvalue; Determine the maximum eigenvalue corresponding to each of the sample matrices according to the eigenvalue information, and determine the radius of the Brauer disk based on the minimum value of each of the maximum eigenvalues; The classification limit is determined by calculation based on the arithmetic mean of the centers of the largest Brauer disks, the geometric mean of the centers of the largest Brauer disks, and the radius of the Brauer disks.

[0010] According to a sea clutter suppression method provided by the present invention, the step of screening the sample matrix according to the classification boundary to obtain the clutter matrix includes: The maximum Brauer disk center is compared with the classification limit, and the sample matrix in which the maximum Brauer disk center is located within the classification limit is determined as the clutter matrix.

[0011] According to a sea clutter suppression method provided by the present invention, constructing a Grassmann manifold according to each of the eigenvector matrices and the corresponding eigenvalue diagonal matrix includes: According to the eigenvalue diagonal matrix corresponding to each of the clutter matrices, based on the order of eigenvalues ​​from large to small, selecting a first preset number of target eigenvalues; Combining the eigenvectors corresponding to the target eigenvalues ​​in the eigenvector matrix to form a subspace corresponding to the clutter matrix; The subspaces corresponding to the clutter matrices form a set to constitute the Grassmann manifold.

[0012] According to a sea clutter suppression method provided by the present invention, the clutter covariance matrix is ​​obtained by calculating based on the Grassmann manifold and each of the clutter matrices based on a volume correlation function, including: Initializing an estimated covariance matrix, wherein the estimated covariance matrix is ​​within the Grassmann manifold; According to the clutter matrix and a preset volume correlation function model, the estimated covariance matrix is ​​iteratively updated based on a gradient descent method until a preset termination condition is met, and the estimated covariance matrix is ​​used as the clutter covariance matrix.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, a sea clutter suppression method as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the sea clutter suppression method as described in any one of the above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned sea clutter suppression methods.

[0016] The present invention provides a method, device and storage medium for suppressing sea clutter, which at least have beneficial effects: determining training data in radar echo data, determining a corresponding sample matrix for each training sample in the training data, setting the sample matrix to be a Toeplitz-Hermitt positive definite matrix, and assuming that the clutter conforms to a specific distribution in order to improve the detection performance. According to the Brauer disk theorem, the classification boundary is determined, and then the sample matrix is ​​screened to screen out the sample matrix containing the target signal, and the clutter matrix is ​​obtained to avoid the influence of the target signal in the subsequent process of obtaining the clutter covariance matrix, and prevent the self-zero effect. According to the clutter matrix, a subspace manifold is constructed to utilize the geometric characteristics of the matrix through the subspace manifold, estimate and obtain the clutter covariance matrix, and achieve the effect of utilizing the structural information of the data. According to the clutter covariance matrix, the space-time filter weight vector is determined so that the space-time filter can accurately suppress sea clutter. Therefore, the clutter matrix corresponding to the non-target signal is screened based on the Brauer disk theorem to adapt to the situation where there are targets with extended range, avoid the self-zero effect and improve the accuracy of sea clutter suppression. At the same time, based on the construction of subspace manifolds, the structured information of the data can be fully utilized to further improve the accuracy of sea clutter suppression. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of a sea clutter suppression method provided by the present invention.

[0019] Figure 2 It is a schematic diagram of the processing process of a sea clutter suppression method provided by the present invention.

[0020] Figure 3 It is a schematic diagram of classification boundaries in one embodiment of a sea clutter suppression method provided by the present invention.

[0021] Figure 4 It is the space-time spectrum of the simulation.

[0022] Figure 5 It is a diagram showing the inhibition results of STAP in the related art.

[0023] Figure 6This is a suppression result diagram of a sea clutter suppression method provided by the present invention.

[0024] Figure 7 It is a comparison diagram of clutter suppression improvement factors of a sea clutter suppression method provided by the present invention and other methods.

[0025] Figure 8 This is a clutter suppression comparison diagram of a sea clutter suppression method provided by the present invention and other methods under different signal-to-noise ratios.

[0026] Fig. 9 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Combine the following Figures 1 to 3 A sea clutter suppression method of the present invention is described, comprising: S100: Acquire radar echo data and determine training data, where the training data includes a plurality of training samples; S200: Acquire a sample matrix corresponding to the training sample according to the training data, where the sample matrix is ​​a Toeplitz-Hermitian positive definite matrix; S300: determining a classification boundary according to the sample matrix and based on the Brauer disk theorem, wherein the classification boundary represents a boundary between a target signal and a clutter signal; S400: Filtering the sample matrix according to the classification boundary to obtain a clutter matrix; S500: Acquire a clutter covariance matrix based on constructing a subspace manifold according to each of the clutter matrices; S600: Determine a space-time filter weight vector according to the clutter covariance matrix.

[0029] Determine the training data in the radar echo data, determine the corresponding sample matrix for each training sample in the training data, let the sample matrix be a Toeplitz-Hermitt positive definite matrix, and assume that the clutter conforms to a specific distribution to improve the detection performance. According to the Brauer disk theorem, determine the classification boundary, and then screen the sample matrix to screen out the sample matrix containing the target signal, obtain the clutter matrix, avoid the influence of the target signal in the subsequent acquisition of the clutter covariance matrix, and prevent the self-zero effect. According to the clutter matrix, construct a subspace manifold to utilize the geometric characteristics of the matrix through the subspace manifold, estimate the clutter covariance matrix, and achieve the effect of utilizing the structural information of the data. According to the clutter covariance matrix, determine the space-time filter weight vector so that the space-time filter can accurately suppress sea clutter.

[0030] Therefore, the clutter matrix corresponding to the non-target signal is screened based on the Brauer disk theorem to adapt to the situation where there are targets with extended range, avoid the self-zero effect and improve the accuracy of sea clutter suppression. At the same time, based on the construction of subspace manifolds, the structured information of the data can be fully utilized to further improve the accuracy of sea clutter suppression.

[0031] The space-time filter weight vector is used to determine the parameters of the space-time two-dimensional filter, so that the space-time two-dimensional filter can adaptively and accurately suppress sea clutter.

[0032] By assuming that the clutter covariance matrix has a specific distribution, the radar detection performance can be improved by smoothing the estimation and reducing the variance. Therefore, the sample matrix is ​​set to be a Toeplitz-Hermitian PositiveDefinite (THPD) matrix, and the clutter covariance matrix is ​​also a THPD matrix, which is beneficial to improving the radar detection performance. The THPD matrix has the characteristics that the elements on each diagonal of the matrix are the same, the conjugate transpose of the matrix is ​​equal to itself, and all the eigenvalues ​​of the matrix are positive. Sea clutter has temporal and spatial correlation, and this correlation can be naturally represented by the Toeplitz structure. The obtained clutter covariance matrix is ​​more consistent with the actual signal model, which helps to improve the detection and estimation performance. The use of the THPD matrix can effectively utilize the sparsity and prior information of the signal, and improve the estimation accuracy of the covariance matrix and the performance of the algorithm.

[0033] In the related art, the clutter covariance matrix estimation is based on the Euclidean metric, but the positive definiteness of the matrix is ​​not always maintained under addition and scalar multiplication, the effect is not ideal when processing the THPD matrix, and the intrinsic geometric properties and structural characteristics of the covariance matrix are ignored, and the structural information of the data cannot be fully utilized. Specifically, the traditional clutter covariance matrix estimation is based on the Euclidean metric, and the Euclidean metric does not always maintain the positive definiteness of the matrix under addition and scalar multiplication. For example, when a positive definite matrix is ​​multiplied by a negative scalar, the result is negative definite, not positive definite, and negative eigenvalues ​​may appear, which makes it undesirable when processing the THPD matrix, because positive definiteness is a property that must be maintained. Therefore, it is inappropriate to consider the covariance matrix problem in Euclidean space, and the intrinsic geometric properties and structural characteristics of the covariance matrix are ignored, and the structural information of the data cannot be fully utilized.

[0034] In the present invention, by constructing a subspace manifold and performing calculation processing on the basis of the subspace manifold, not only can the positive definiteness of the matrix be maintained, but also the structural information of the geometric characteristics of the data can be fully utilized, thereby making the estimated clutter covariance matrix have higher accuracy.

[0035] In some embodiments of a method for suppressing sea clutter of the present invention, the step S500 includes: S510: Decomposing the clutter matrix to obtain a corresponding eigenvector matrix and an eigenvalue diagonal matrix; S520: constructing a Grassmann manifold according to each of the eigenvector matrices and the corresponding eigenvalue diagonal matrix; S530: performing calculations based on the volume correlation function according to the Grassmann manifold and each of the clutter matrices to obtain the clutter covariance matrix.

[0036] According to the eigenvector matrix and eigenvalue diagonal matrix of each clutter matrix, the Grassmann manifold is constructed, and the estimation of the clutter covariance matrix is ​​realized through the volume correlation function (VCF) based on the Grassmann manifold and the clutter matrix. The data from the Euclidean space is represented as a subspace of fixed dimension through the Grassmann manifold to capture the angle, direction and other non-Euclidean properties in the data, so as to make full use of the structural information of the data for the purpose of estimating the clutter covariance matrix, which is conducive to improving the accuracy of the clutter covariance matrix.

[0037] In some embodiments of the present invention, the subspace manifold may be an implementation of other manifolds besides the Grassmann manifold.

[0038] refer to Figure 2In some embodiments of a method for suppressing sea clutter of the present invention, the step S100 includes: Acquire radar echo data, wherein the radar echo data includes data of array element dimension, pulse dimension and distance dimension; According to the radar echo data, based on the position of the sliding window in the distance dimension, determining a preset number of the training samples to form the training data; The length of the sliding window is a preset value, the sliding window is used to move along the distance dimension and determine the range of data processing according to the position of the sliding window, and the training sample includes a matrix formed by corresponding array element dimension data and pulse dimension data at the distance.

[0039] The radar system consists of multiple array elements, which emit a fixed number of pulses in each coherent processing interval (CPI). When acquiring echoes, multiple distance units are used to distinguish the distances. Figure 2 The radar echo data includes data of array element dimension, pulse dimension and distance dimension. In order to detect whether there is a target signal in the distance unit, the distance unit is used as the unit to be tested. In order to suppress sea clutter, the clutter covariance matrix needs to be estimated from other distance units. At the same time, several distance units adjacent to the unit to be tested are used as protection units. The unit to be tested is located at the center of the sliding window. The distance unit used as the training sample is determined based on the position and length of the sliding window.

[0040] It can be understood that the sliding window moves along the distance dimension to use different distance units as units to be tested to determine whether the unit to be tested includes the target signal.

[0041] For ease of understanding, a schematic example is given, where the radar echo data includes N array element dimension data, M pulse dimension data, and K distance dimension data, and the distance unit includes the array element dimension data and the pulse dimension data at the same distance, then the data included in the kth distance unit can be expressed as: K range units expand the range dimension, and the radar echo data can be expressed as: right Perform vectorization processing to obtain the space-time snapshots of K distance units: In short: .

[0042] For each distance unit as a unit to be tested, the unit to be tested and several adjacent distance units as protection units are removed based on the sliding window, and other distance units are used as training samples. According to the sliding window assumption, there are Q training samples determined, and the qth training sample can be expressed as: .

[0043] Let the sample matrix be the THPD matrix, that is, the sample matrix corresponding to the training sample is: The reflection coefficient Calculate the following expression: in, and reflection coefficient Obtained through the regularized Burg algorithm, and The values ​​are as follows: refer to Figure 3 In some embodiments of a method for suppressing sea clutter of the present invention, the step S300 includes: Performing eigenvalue decomposition on the sample matrix, obtaining corresponding eigenvalue information and determining the maximum Brauer disk center, wherein the maximum Brauer disk center is determined by the maximum eigenvalue; Determine the maximum eigenvalue corresponding to each of the sample matrices according to the eigenvalue information, and determine the radius of the Brauer disk based on the minimum value of each of the maximum eigenvalues; The classification limit is determined by calculation based on the arithmetic mean of the centers of the largest Brauer disks, the geometric mean of the centers of the largest Brauer disks, and the radius of the Brauer disks.

[0044] Perform eigenvalue decomposition on the sample matrix to obtain the eigenvalue information of the sample matrix. According to the Brauer's Oval Theorem, determine the center of the maximum Brauer disk, that is, the area containing the maximum eigenvalue. According to the eigenvalue information of each sample matrix, based on the maximum eigenvalue of each sample matrix, determine the smallest maximum eigenvalue, and calculate the radius of the Brauer disk. The radius of the Brauer disk describes the distribution range of the eigenvalue on the complex plane, reflects the degree of disturbance of the sample matrix, and represents the minimum limit of clutter noise. According to the maximum Brauer disk center and the radius of the Brauer disk of each sample matrix, determine the classification boundary. In this way, due to the difference in the matrix eigenvalues ​​corresponding to the clutter signal and the target signal, the classification boundary is obtained based on the Brauer disk theorem to distinguish the sample matrices corresponding to the clutter signal and the target signal.

[0045] In some embodiments of the present invention, the classification boundary can be obtained by calculating the following expression: in, For classification boundaries; is the center of the maximum Brauer disk of the qth sample matrix; is the arithmetic mean of the centers of the largest Brauer disks; is the geometric mean of the centers of the largest Brauer disks; is the radius of the Brauer disk.

[0046] refer to Figure 3 In some embodiments of a sea clutter suppression method of the present invention, the S400 includes: The maximum Brauer disk center is compared with the classification limit, and the sample matrix in which the maximum Brauer disk center is located within the classification limit is determined as the clutter matrix.

[0047] The maximum Brauer disk center of each sample matrix is ​​compared with the classification boundary. If the maximum Brauer disk center is within the classification boundary, the corresponding sample matrix is ​​considered to be formed by pure clutter signals, and if the maximum Brauer disk center is outside the classification boundary, the corresponding sample matrix is ​​considered to potentially contain target signals. In this way, the sample matrix is ​​screened based on the comparison results, and the sample matrix corresponding to the clutter signal is determined and used as the clutter matrix, avoiding the use of the sample matrix corresponding to the target signal as the subsequent processing sample, which is conducive to avoiding the self-zero effect and adapting to the situation where there is a distance extension target.

[0048] In some embodiments of a sea clutter suppression method of the present invention, the S520 includes: According to the eigenvalue diagonal matrix corresponding to each of the clutter matrices, based on the order of eigenvalues ​​from large to small, selecting a first preset number of target eigenvalues; Combining the eigenvectors corresponding to the target eigenvalues ​​in the eigenvector matrix to form a subspace corresponding to the clutter matrix; The subspaces corresponding to the clutter matrices form a set to constitute the Grassmann manifold.

[0049] Based on the classification boundary, after screening the Q sample matrices, L clutter matrices are determined. The clutter matrix is ​​decomposed, such as eigenvalue decomposition or singular value decomposition, to obtain the eigenvector matrix and eigenvalue diagonal matrix. Since the clutter matrix is ​​also a THPD matrix, it has the Hermitian positive definite property and can be expressed as: in, For the A clutter matrix; is the eigenvector matrix; is the eigenvalue diagonal matrix; is the transposed conjugate matrix of the eigenvector matrix.

[0050] According to the eigenvalue diagonal matrix , select r eigenvalues ​​from large to small, determine the corresponding r eigenvectors, and form a NM×r matrix corresponding to the clutter matrix ,matrix The column space of is an r-dimensional subspace. The matrix corresponding to each clutter matrix is Set, forming a Grassmann manifold .

[0051] On the Grassmann manifold Based on the calculation processing, the characteristics of the clutter matrix are mapped to the Grassmann manifold , which not only maintains the positive definiteness of the matrix, but also makes full use of the structural information of the geometric characteristics of the data, overcomes the defects of the Euclidean metric, and makes the estimated clutter covariance matrix more accurate.

[0052] In some embodiments of a method for suppressing sea clutter of the present invention, the step S530 includes: Initializing an estimated covariance matrix, wherein the estimated covariance matrix is ​​within the Grassmann manifold; According to the clutter matrix and a preset volume correlation function model, the estimated covariance matrix is ​​iteratively updated based on a gradient descent method until a preset termination condition is met, and the estimated covariance matrix is ​​used as the clutter covariance matrix.

[0053] Based on Grassmann manifold Initialize the estimated covariance matrix, then use the clutter matrix to iteratively update the estimated covariance matrix through the preset volume correlation function model, and continuously optimize and update the estimated covariance matrix based on the gradient descent method until the preset termination condition is met. The final estimated covariance matrix is ​​used as the estimated clutter covariance matrix. The volume correlation function (VCF) and gradient descent method are used to estimate the clutter covariance matrix. The matrix positive definiteness is maintained during the iterative update process. The structural information of the geometric characteristics of the data is fully utilized to accurately estimate the clutter covariance matrix, which is conducive to minimizing the power of the clutter signal without losing the power of the target signal, thereby achieving the purpose of suppressing sea clutter.

[0054] The estimated covariance matrix is , the estimated covariance matrix can be decomposed into ,in, is the eigenvector matrix of the estimated covariance matrix; is the diagonal matrix of eigenvalues ​​of the estimated covariance matrix; is the transposed conjugate matrix of the eigenvector matrix.

[0055] In some embodiments of the present invention, the preset volume correlation function model may be: in, is the weight parameter; For the The matrix corresponding to the r-dimensional subspace of the clutter matrix; is the matrix corresponding to the estimated covariance matrix r-dimensional subspace; is a Grassmann manifold The volume correlation function in .

[0056] In some embodiments of the present invention, the volume correlation function may specifically be: in, is the matrix volume, denotes the singular values ​​of the corresponding matrix.

[0057] Understandably, The process of obtaining The same, that is, according to the eigenvalue diagonal matrix , select r eigenvalues ​​from large to small and determine The corresponding r eigenvectors in form a corresponding NM×r matrix .

[0058] The objective function model is minimized using the gradient descent method to update the estimated covariance matrix. The iterative update of the estimated covariance matrix can be expressed as: After updating the eigenvectors of the r-dimensional subspace of the estimated covariance matrix, recalculate the updated estimated covariance matrix .

[0059] Check the convergence of the optimization iterative update process. If the termination conditions such as the objective function change being less than the preset threshold are met, stop the iterative update and obtain the estimated covariance matrix. as the clutter covariance matrix.

[0060] The eigenvector matrix and eigenvalue diagonal matrix of the clutter matrix are processed to form the corresponding subspaces. The collection of each subspace constitutes a Grassmann manifold. The volume correlation function VCF is used to measure the geometric correlation between different subspaces on the Grassmann manifold. By minimizing the volume correlation distance between subspaces, the global optimal clutter covariance matrix is ​​obtained through iterative optimization. At the same time, since the clutter matrix is ​​a THPD matrix, the sparsity and prior information of the signal are utilized, and the robust covariance matrix estimation is achieved by directly measuring the distance between the sampled signal subspace and the target subspace, which is conducive to the completion of sea clutter suppression.

[0061] In some embodiments of the present invention, the space-time filter weight vector is calculated according to the clutter covariance matrix, which can be obtained by optimizing the following expression: in, is the normalized Doppler frequency of the target signal; is the normalized spatial frequency of the target signal; t represents the target; is the target orientation vector; is the clutter covariance matrix; is the space-time filter weight vector.

[0062] The space-time filter weight vector can be obtained by the following expression: After obtaining the space-time filter weight vector, a space-time filter is constructed to suppress clutter on the unit under test, which is conducive to accurately detecting the target signal and improving the performance of the radar.

[0063] In order to facilitate understanding of the present solution, one embodiment of the present invention is schematically illustrated: The radar system consists of N = 10 array elements. M = 12 pulses are transmitted in each coherent processing interval. For K = 80 range units, radar echo data is obtained. A sliding window with a preset length of 26 is used to slide the received radar echo data along the range dimension to determine the unit to be tested, the protection unit and Q = 23 training samples. Figure 2 shown.

[0064] For Q=23 training samples, the corresponding sample matrices are calculated respectively, and the sample matrix is ​​the THPD matrix.

[0065] For the sample matrix with Q=23 training samples, based on the Brauer disk theorem, the classification boundary of the sample matrix of clutter and target signal is determined as ,like Figure 3 As shown, based on the classification boundaries As well as the maximum Brauer disk center of Q=23 sample matrices, the sample matrix corresponding to the eigenvalues ​​within the Brauer clustering boundary is regarded as the sample matrix of clutter plus noise, while the sample matrix corresponding to the eigenvalues ​​outside the Brauer cluster boundary is considered to be the sample matrix of potential targets, and L=19 clutter matrices are obtained by screening.

[0066] For L = 19 120 × 120 dimensional clutter matrices , perform eigenvalue decomposition or singular value decomposition, . Select r = 5 eigenvectors (corresponding to the largest 5 eigenvalues) from each eigenvector matrix to form a clutter matrix The corresponding 120×5 matrix . Each clutter matrix Set, forming a Grassmann manifold .

[0067] Based on the obtained Grassmann manifold , initialize the estimated covariance matrix and iteratively update it by presetting the volume correlation function model: Then the estimated covariance matrix obtained by the final optimization is used as the clutter covariance matrix .

[0068] Construct the space-time steering vector according to the clutter covariance matrix Calculate the space-time filter weight vector to achieve sea clutter suppression of the unit under test. The purpose of the space-time filter is to minimize the clutter filter output power while ensuring that the output power at the target position is not lost. Figure 4 The space-time spectrum of the original simulation is shown, Figure 5 It represents the suppression result diagram of the STAP method in the related art, Figure 6The result diagram of the sea clutter suppression method provided by the present invention shows that the sea clutter suppression method provided by the present invention is better than the STAP method in the related art.

[0069] The Improvement Factor (IF) curve is used as a benchmark for evaluating the performance of different algorithms. The IF curve is defined by the following formula: Where R represents the true clutter plus noise covariance matrix; Represents matrix trace calculation; is the normalized Doppler frequency of the target signal; is the normalized spatial frequency of the target signal; t represents the target; is the target orientation vector; is the space-time filter weight vector.

[0070] like Figure 7 As shown in the figure, under the same input signal-to-noise ratio, the improvement factor comparison diagram of other methods and the method provided by the present invention shows that the performance of the sea clutter suppression method BGVCF-STAP provided by the present invention is significantly better than other methods by about 5dB, that is, it can improve the signal-to-noise ratio to a greater extent.

[0071] like Figure 8 The clutter suppression performance comparison diagram of other methods and the sea clutter suppression method provided by the present invention under different signal-to-noise ratios shows that under different signal-to-noise ratios, the performance of the method BGVCF-STAP of the present invention is still better than other methods.

[0072] Fig. 9 An example of a physical structure diagram of an electronic device is shown in FIG. Fig. 9 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820 and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the above-mentioned sea clutter suppression method.

[0073] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0074] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a sea clutter suppression method provided by the above methods.

[0075] In yet another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute a sea clutter suppression method provided by the above methods.

[0076] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0077] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0078] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0079] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0080] All actions to obtain signals, information or data in this application are carried out in compliance with the relevant data protection laws and policies of the location and with the authorization given by the owner of the corresponding device.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A sea clutter suppression method, characterized in that: include: Acquire radar echo data and determine training data, wherein the training data includes a plurality of training samples; According to the training data, a sample matrix corresponding to the training sample is obtained, where the sample matrix is ​​a Toeplitz-Hermitian positive definite matrix; According to the sample matrix, based on the Brauer disk theorem, a classification boundary is determined, wherein the classification boundary represents a boundary between a target signal and a clutter signal; Filter the sample matrix according to the classification boundary to obtain a clutter matrix; According to each of the clutter matrices, based on constructing a subspace manifold, obtaining a clutter covariance matrix; A space-time filter weight vector is determined according to the clutter covariance matrix.

2. A sea clutter suppression method according to claim 1, characterized in that: The step of obtaining a clutter covariance matrix based on constructing a subspace manifold according to each of the clutter matrices includes: Decomposing the clutter matrix to obtain a corresponding eigenvector matrix and an eigenvalue diagonal matrix; Constructing a Grassmann manifold according to each of the eigenvector matrices and the corresponding eigenvalue diagonal matrices; The clutter covariance matrix is ​​obtained by performing calculations based on the Grassmann manifold and each of the clutter matrices and based on a volume correlation function.

3. A sea clutter suppression method according to claim 2, characterized in that: The step of acquiring radar echo data and determining training data includes: Acquire radar echo data, wherein the radar echo data includes data of array element dimension, pulse dimension and distance dimension; According to the radar echo data, based on the position of the sliding window in the distance dimension, determining a preset number of the training samples to form the training data; The length of the sliding window is a preset value, the sliding window is used to move along the distance dimension and determine the range of data processing according to the position of the sliding window, and the training sample includes a matrix formed by corresponding array element dimension data and pulse dimension data at the distance.

4. A sea clutter suppression method according to claim 2, characterized in that: Determining the classification boundary according to the sample matrix based on the Brauer disk theorem includes: Performing eigenvalue decomposition on the sample matrix, obtaining corresponding eigenvalue information and determining the maximum Brauer disk center, wherein the maximum Brauer disk center is determined by the maximum eigenvalue; Determine the maximum eigenvalue corresponding to each of the sample matrices according to the eigenvalue information, and determine the radius of the Brauer disk based on the minimum value of each of the maximum eigenvalues; The classification limit is determined by calculation based on the arithmetic mean of the centers of the largest Brauer disks, the geometric mean of the centers of the largest Brauer disks, and the radius of the Brauer disks.

5. A sea clutter suppression method according to claim 4, characterized in that: The step of screening the sample matrix according to the classification boundary to obtain a clutter matrix includes: The maximum Brauer disk center is compared with the classification limit, and the sample matrix in which the maximum Brauer disk center is located within the classification limit is determined as the clutter matrix.

6. A sea clutter suppression method according to claim 2, characterized in that: The step of constructing a Grassmann manifold according to each of the eigenvector matrices and the corresponding eigenvalue diagonal matrix comprises: According to the eigenvalue diagonal matrix corresponding to each of the clutter matrices, based on the order of eigenvalues ​​from large to small, selecting a first preset number of target eigenvalues; Combining the eigenvectors corresponding to the target eigenvalues ​​in the eigenvector matrix to form a subspace corresponding to the clutter matrix; The subspaces corresponding to the clutter matrices form a set to constitute the Grassmann manifold.

7. A sea clutter suppression method according to claim 2, characterized in that: The step of calculating according to the Grassmann manifold and each of the clutter matrices based on a volume correlation function to obtain the clutter covariance matrix includes: Initializing an estimated covariance matrix, wherein the estimated covariance matrix is ​​within the Grassmann manifold; According to the clutter matrix and a preset volume correlation function model, the estimated covariance matrix is ​​iteratively updated based on a gradient descent method until a preset termination condition is met, and the estimated covariance matrix is ​​used as the clutter covariance matrix.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the sea clutter suppression method according to any one of claims 1 to 7 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, a sea clutter suppression method as claimed in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, a sea clutter suppression method as claimed in any one of claims 1 to 7 is implemented.

Images