Semi-parametric fast suppression method for radio frequency interference in SAR system
By using low-rank methods to stitch and suppress multi-polarization SAR data matrices, the problem of difficult removal of radio frequency interference in polarization SAR systems is solved, achieving more effective interference suppression and real echo protection, and simplifying computational complexity.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SOUTHEAST UNIV
- Filing Date
- 2023-06-29
- Publication Date
- 2026-05-15
AI Technical Summary
Existing technologies struggle to effectively remove radio frequency interference in polarimetric SAR systems, especially in low-rank methods where multi-polarization data fails to effectively aid interference suppression and computational complexity is high.
By performing matrix stitching and interference suppression on the multipolar SAR data matrix using low-rank methods, and constructing a low-rank sparse model for interference suppression using principal component analysis, robust principal component analysis, or complex tensor robust principal component analysis, the multipolar data is recovered through the inverse process.
It improves the suppression of radio frequency interference in polarimetric SAR systems, enhances the protection of real echo signals, and simplifies the calculation process.
Smart Images
Figure CN116819450B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of polarimetric SAR data processing technology, and in particular to a semi-parametric fast suppression method for radio frequency interference in SAR systems. Background Technology
[0002] Synthetic Aperture Radar (SAR) is a microwave sensor capable of providing high-resolution images of observed scenes under various weather and lighting conditions. Polarimetric SAR, in particular, can provide multi-polarization data and is therefore widely used in SAR missions. However, due to increasingly congested spectrums, polarimetric SAR systems may be subject to interference from other systems during imaging; this interference is known as radio frequency (RF) interference. Therefore, how to remove or mitigate RF interference in polarimetric SAR systems has attracted considerable attention from the academic community.
[0003] Currently, traditional methods for removing radio frequency interference from single-polarization SAR data include nonparametric, parametric, and semiparametric methods.
[0004] While traditional nonparametric methods are simple and fast and do not depend on the waveform characteristics of interference, they may fail when the interference is weak. Therefore, such algorithms are not robust enough in time-varying and complex interference environments. At the same time, the signal actually received by radar is a combination of interference, real echo, and noise. Nonparametric algorithms only consider the intensity characteristics of interference and lack constraints on the real echo, thus failing to effectively protect it.
[0005] Parametric methods can remove interference relatively well, but they require the interference to be isolated and of a single type, and they have high computational complexity. They also do not take into account the characteristics of the real echo signal and lack protection for it.
[0006] For semi-parametric methods, not only do they use optimization models to constrain interference, but they also take into account the characteristics of the real echo signal. However, the suppression effect is highly dependent on the established model, and the computational load is huge when performing large-scale matrix operations.
[0007] Traditional low-rank methods are generally ineffective against radio frequency interference, and they all use only single polarization data for interference suppression, rarely utilizing different polarization data for the same scenario for auxiliary interference suppression. Summary of the Invention
[0008] This invention provides a semi-parametric fast suppression method for radio frequency interference in SAR systems, in order to solve the problem that there is a large amount of radio frequency interference in current polarimetric SAR systems, and low-rank methods are unable to remove some of the radio frequency interference.
[0009] This invention provides a semi-parametric fast suppression method for radio frequency interference in SAR systems, comprising the following steps:
[0010] Obtain the N polarization SAR data matrices of the SAR system that are interfered with at the same time and in the same scene, where N is greater than or equal to 2 and less than or equal to 4;
[0011] The N polarization SAR data matrices are stitched together using a preset low-rank class method to obtain a SAR data stitching matrix;
[0012] The SAR data stitching matrix is subjected to interference suppression using the preset low-rank class method to obtain the suppressed SAR data stitching matrix;
[0013] The suppressed SAR data stitching matrix is decomposed using the inverse process of the preset low-rank class method to obtain N polarization SAR data matrices after interference suppression.
[0014] Extract the SAR data matrix of any polarization to obtain the SAR image after removing interference. Among them, the SAR data matrices other than the SAR data matrix of any polarization contain data with greater interference than the SAR data matrix of any polarization.
[0015] Optionally, in one embodiment of the present invention, the low-rank class method includes at least one of principal component analysis, robust principal component analysis, and complex tensor robust principal component analysis.
[0016] Optionally, in one embodiment of the present invention, when using the low-rank method as the principal component analysis method and the robust principal component analysis method, a preset low-rank method is used to perform matrix stitching on the N polarization SAR data matrices to obtain a SAR data stitching matrix, including:
[0017] When N≤3, if k>l, then the SAR data stitching matrix is [X1,X2,…,X]. N If k≤l, then the SAR data stitching matrix is [X1; X2; ...; X...]. N ], where k is the number of rows in the SAR data matrix and l is the number of columns in the SAR data matrix;
[0018] When N=4, if k>4l, the SAR data stitching matrix is [X1,X2,X3,X4], if 4k≤l, the SAR data stitching matrix is [X1;X2;X3;X4], and if k≤4l or 4k>l, the SAR data stitching matrix is [X1,X2;X3,X4].
[0019] Optionally, in one embodiment of the present invention, when the low-rank method is a complex tensor robust principal component analysis method, a preset low-rank method is used to perform matrix concatenation on the N polarization SAR data matrices to obtain a SAR data concatenation matrix, including:
[0020] Divide the number of rows k of each polarization SAR data matrix into p parts to obtain p SAR data segmentation matrices Z1, Z2, ..., Zn corresponding to each polarization SAR data matrix. P For each polarization of the SAR data matrix, the same part Z of the SAR data cutting matrix is... i1 Z i2 ,…,Z iN To splice them together, Z iN This represents the Nth polarization SAR data matrix of the i-th block, with each matrix having a size of S*l;
[0021] When N≤3, if S>1, then the SAR data stitching matrix is [Z i1 Z i2 ,…,Z iN If S≤l, then the SAR data stitching matrix is [Z]. i1 Z i2 ;…;Z iN ];
[0022] When N=4, if S>4l, then the SAR data stitching matrix is [Z i1 Z i2 Z i3 Z i4 If 4S≤l, then the SAR data stitching matrix is [Z]. i1 Z i2 Z i3 Z i4 If S≤4l and 4S>l, then the SAR data stitching matrix is [Z i1 Z i2 Z i3 Z i4 ].
[0023] The semi-parametric fast suppression method for radio frequency interference in SAR systems according to embodiments of the present invention processes multi-polarization SAR data using a low-rank class method, and ultimately achieves the auxiliary suppression of radio frequency interference by multi-polarization data in polarimetric SAR systems, thereby improving the interference suppression effect.
[0024] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description
[0025] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:
[0026] Figure 1 A flowchart of a semi-parametric fast suppression method for radio frequency interference in a SAR system according to an embodiment of the present invention;
[0027] Figure 2 The original Sentinel-1A VV polarization image provided according to an embodiment of the present invention;
[0028] Figure 3 This is the original Sentinel-1A VH polarization image provided according to an embodiment of the present invention;
[0029] Figure 4 The splicing process provided according to an embodiment of the present invention;
[0030] Figure 5 The image is a Sentinel-1A VV polarization original image provided according to an embodiment of the present invention, processed by the present invention. Detailed Implementation
[0031] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.
[0032] The following describes a semi-parametric fast suppression method for radio frequency interference (RF) in SAR systems according to embodiments of the present invention, with reference to the accompanying drawings. Addressing the problem mentioned in the background section that current polarimetric SAR systems suffer from significant RF interference, and that low-rank methods struggle to remove some of this interference, the present invention provides a semi-parametric fast suppression method for RF interference in SAR systems. In this method, a multi-polarimetric SAR data matrix is received; the multi-polarimetric matrix is stitched together based on a selected low-rank interference suppression method; interference suppression is applied to the stitched data; and the interference-suppressed data is restored to its original form, recovering multiple polarimetric matrices. This method can be directly applied to raw SAR system data or SAR images, improving the SAR interference suppression effect of low-rank methods.
[0033] Specifically, Figure 1 This is a flowchart of a semi-parametric fast suppression method for radio frequency interference in a SAR system according to an embodiment of the present invention.
[0034] like Figure 1As shown, the semi-parametric fast suppression method for radio frequency interference in SAR systems includes the following steps:
[0035] In step S101, the N polarization SAR data matrices of the SAR system that are interfered with at the same time and in the same scene are obtained, where N is greater than or equal to 2 and less than or equal to 4.
[0036] In an embodiment of the present invention, for a specific interfering polarization data to be suppressed, firstly, a polarization matrix data X1 of a polarimetric SAR image is received, and then different polarization data X2, ..., X2 received by the SAR system at the same time and in the same scene are selected. N The matrix contains data with stronger interference than X1.
[0037] Where X1, X2, ..., X N It contains various polarization data, with N being at most 4, and a maximum of four polarization data types, namely VV, VH, HV, and HH. VV is vertical polarization transmission and vertical polarization reception, VH is vertical polarization transmission and horizontal polarization reception, HV is horizontal polarization transmission and vertical polarization reception, and HH is horizontal polarization transmission and horizontal polarization reception.
[0038] In step S102, the N polarization SAR data matrices are stitched together using a preset low-rank class method to obtain a SAR data stitching matrix.
[0039] Based on the above embodiments, a certain low-rank class method is selected, and for this class of methods, X1, X2, ..., X N Matrix concatenation is performed, where low-rank methods include, but are not limited to, Principal Component Analysis (PCA), Robust Principal Component Analysis (RPCA), and Complex Tensor Robust Principal Component Analysis (CT-RPCA). The concatenation method involves combining multiple polarization matrices into a single matrix according to the method's requirements.
[0040] In one embodiment of the present invention, for low-rank class methods such as PCA and RPCA that do not require changing the shape of the original matrix, a preferred splicing method is that the spliced matrix only needs to determine the single polarization matrix X. k*l The size, where k is the number of rows in matrix X, which is also the number of pulses, and l is the number of columns in matrix X.
[0041] When using the low-rank method as the principal component analysis method and the robust principal component analysis method, a pre-defined low-rank method is used to perform matrix stitching on SAR data matrices of N polarizations to obtain a SAR data stitching matrix, including:
[0042] When N≤3, if k>l, then the SAR data stitching matrix is [X1,X2,…,X]. N If k≤l, then the SAR data stitching matrix is [X1; X2; ...; X...].N ], where k is the number of rows in the SAR data matrix and l is the number of columns in the SAR data matrix;
[0043] When N=4, if k>4l, the SAR data stitching matrix is [X1,X2,X3,X4], if 4k≤l, the SAR data stitching matrix is [X1;X2;X3;X4], and if k≤4l or 4k>l, the SAR data stitching matrix is [X1,X2;X3,X4].
[0044] In an embodiment of the present invention, when the low-rank method is a complex tensor robust principal component analysis method, a preset low-rank method is used to perform matrix stitching on SAR data matrices of N polarizations to obtain a SAR data stitching matrix, including:
[0045] Divide the number of rows k of each polarization SAR data matrix into p parts to obtain p SAR data segmentation matrices Z1, Z2, ..., Zn corresponding to each polarization SAR data matrix. P For each polarization of the SAR data matrix, the same part Z of the SAR data cutting matrix is... i1 Z i2 ,…,Z iN To splice them together, Z iN This represents the Nth polarization SAR data matrix of the i-th block, with each matrix having a size of S*l;
[0046] Specifically, the CT-RPCA method selects the number of viewing angles p, divides the original matrix into p parts based on the number of pulses k, and each resulting matrix consists of s = k / p pulses. That is, each polarization matrix is cut into new matrices Z1, Z2, ..., Z... P .
[0047] When N≤3, if S>l, then the SAR data stitching matrix is [Z i1 Z i2 ,…,Z iN If S≤l, then the SAR data stitching matrix is [Z]. i1 Z i2 ;…;Z iN ];
[0048] When N=4, if S>4l, then the SAR data stitching matrix is [Z i1 Z i2 Z i3 Z i4 If 4S≤l, then the SAR data stitching matrix is [Z]. i1 Z i2 Z i3 Z i4 If S≤4l and 4S>l, then the SAR data stitching matrix is [Zi1 Z i2 Z i3 Z i4 ].
[0049] In step S103, interference suppression is performed on the SAR data stitching matrix using a preset low-rank method to obtain the suppressed SAR data stitching matrix.
[0050] The low-rank class method selected in the above embodiments is used to suppress interference in the spliced matrix or tensor.
[0051] In one embodiment of the present invention, a preset low-rank class method is used to suppress interference in the SAR data stitching matrix.
[0052] In the selected low-rank class method, when using principal component analysis, singular value decomposition is performed on the stitched matrix to obtain the left singular vector, singular value matrix, and right singular vector. By selecting an appropriate threshold, the singular value part in the singular value matrix that is greater than the threshold is regarded as the interference part, and the remaining singular value part is regarded as the signal part. Then, the extracted singular value matrix and the left and right singular value vectors are reconstructed into a new interference matrix and signal matrix, thereby obtaining the suppressed SAR data stitching matrix.
[0053] When the selected low-rank method is robust principal component analysis, the stitched matrix is constructed as a low-rank sparse model, with interference considered as low-rank terms and signals as sparse terms. The matrix is iteratively solved using the low-rank sparse decomposition method to obtain interference in low-rank terms and signals in sparse terms, thus obtaining the suppressed SAR data stitched matrix.
[0054] When the selected low-rank method is complex tensor robust principal component analysis, the stitched tensor is constructed as a low-rank sparse model, with interference considered as low-rank terms and signals as sparse terms. Low-rank sparse decomposition is performed under the tensor model, and the tensor is iteratively solved to obtain interference in low-rank terms and signals in sparse terms, thereby obtaining the suppressed SAR data stitching matrix.
[0055] Step S104: The suppressed SAR data stitching matrix is decomposed using the inverse process of the preset low-rank class method to obtain N polarization SAR data matrices after interference suppression.
[0056] Step S105: Extract the SAR data matrix of any polarization to obtain the SAR image after removing interference. Among them, the SAR data matrices other than the SAR data matrix of any polarization contain data with greater interference than the SAR data matrix of any polarization.
[0057] Step S102 involves stitching together the SAR data matrix, restoring the matrix or tensor after interference suppression, and reversing the stitching process as described in the above embodiment to recover multiple polarization matrices Y1, Y2, ..., Y N Extracting the Y1 data yields the SAR image after interference removal.
[0058] The following detailed description of the semi-parametric fast suppression method for radio frequency interference in SAR systems according to the present invention is provided through specific embodiments.
[0059] like Figure 2 and Figure 3 As shown, using Sentinel-1A SAR image data, the present invention provides a semi-parametric fast suppression method for radio frequency interference in SAR systems, comprising the following steps:
[0060] Step 1: Use MATLAB software to read the raw VV polarization data of the Sentinel-1A SAR image. The effect of the raw VV polarization image is as follows: Figure 2 As shown.
[0061] Step 2: Read VH polarization data recorded at the same time in the same scene. Radio frequency interference also exists in the VH polarization data, and the interference from VV polarization is weaker than that from VH polarization; the original VH polarization image is shown below. Figure 3 As shown.
[0062] Step 3: Select the RPCA method for interference suppression, and simultaneously splice the VV polarization and VH polarization matrices from Step 2.
[0063] The size of a single polarization matrix is 1400*4096, therefore the VV polarization data and VH polarization data are concatenated to form X = [X VV ,X VH The splicing process is as follows: Figure 4 As shown.
[0064] Step 4: Use the RPCA method to suppress interference in the spliced matrix X.
[0065] Step 5: Restore the matrix after interference suppression by reversing the previous splicing method to recover two polarization matrices Y. VV ,Y VH , where Y VV This is the SAR image after interference removal. The image Y after processing by this invention... VV like Figure 5 As shown.
[0066] The semi-parametric fast suppression method for radio frequency interference in SAR systems proposed in this invention uses a low-rank class method to process multi-polarization SAR data, ultimately achieving multi-polarization data-assisted suppression of radio frequency interference in polarimetric SAR systems, improving the interference suppression effect, and enabling the suppression of radio frequency interference in SAR systems to protect real-world scenarios.
[0067] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0068] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0069] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.
Claims
1. A semi-parametric fast suppression method for radio frequency interference in SAR systems, characterized in that, Includes the following steps: Obtain the N polarization SAR data matrices of the SAR system that are interfered with at the same time and in the same scene, where N is greater than or equal to 2 and less than or equal to 4; The N polarization SAR data matrices are stitched together using a preset low-rank class method to obtain a SAR data stitching matrix; The SAR data stitching matrix is subjected to interference suppression using the preset low-rank class method to obtain the suppressed SAR data stitching matrix; The suppressed SAR data stitching matrix is decomposed using the inverse process of the preset low-rank class method to obtain N polarization SAR data matrices after interference suppression. Extract the SAR data matrix of any polarization to obtain the SAR image after removing interference. Among them, the SAR data matrices other than the SAR data matrix of any polarization contain data with greater interference than the SAR data matrix of any polarization.
2. The method according to claim 1, characterized in that, The low-rank class method includes at least one of principal component analysis, robust principal component analysis, and complex tensor robust principal component analysis.
3. The method according to claim 2, characterized in that, When using the low-rank method for principal component analysis and robust principal component analysis, a preset low-rank method is used to perform matrix stitching on the N polarization SAR data matrices to obtain a SAR data stitching matrix, including: When N≤3, if k>l, then the SAR data stitching matrix is [X1,X2,…,X]. N If k≤l, then the SAR data stitching matrix is [X1; X2; ...; X...]. N ], where k is the number of rows in the SAR data matrix and l is the number of columns in the SAR data matrix; When N=4, if k>4l, the SAR data stitching matrix is [X1,X2,X3,X4], if 4k≤l, the SAR data stitching matrix is [X1;X2;X3;X4], and if k≤4l or 4k>l, the SAR data stitching matrix is [X1,X2;X3,X4].
4. The method according to claim 2, characterized in that, When the low-rank method is a complex tensor robust principal component analysis method, the N polarization SAR data matrices are stitched together using a preset low-rank method to obtain a SAR data stitching matrix, including: Divide the number of rows k of each polarization SAR data matrix into p parts to obtain p SAR data segmentation matrices Z1, Z2, ..., Zn corresponding to each polarization SAR data matrix. P For each polarization of the SAR data matrix, the same part Z of the SAR data cutting matrix is... i1 Z i2 ,…,Z iN To splice them together, Z iN This represents the Nth polarization SAR data matrix of the i-th block, with each matrix having a size of S*l; When N≤3, if S>1, then the SAR data stitching matrix is [Z i1 Z i2 ,…,Z iN If S≤l, then the SAR data stitching matrix is [Z]. i1 Z i2 ;…;Z iN ]; When N=4, if S>4l, then the SAR data stitching matrix is [Z i1 Z i2 Z i3 Z i4 If 4S≤l, then the SAR data stitching matrix is [Z]. i1 Z i2 Z i3 Z i4 If S≤4l and 4S>l, then the SAR data stitching matrix is [Z i1 Z i2 ; WITH i3 ,WITH i4 ]。