High-resolution SAR (Synthetic Aperture Radar) sidelobe suppression method and device based on compressed sensing

Abstract

The invention provides a high-resolution SAR (Synthetic Aperture Radar) sidelobe suppression method and device based on compressed sensing. The method comprises the steps of obtaining a to-be-processed SAR image, and determining a weighting function based on the to-be-processed SAR image and an SVA algorithm; generating a filtered SAR image by using the weighting function; obtaining a measurement matrix, and cutting the measurement matrix by using the filtered SAR image to obtain a measurement cutting matrix; performing main lobe value recovery processing on the filtered SAR image by adopting Moore-Penrose inverse and a measurement cutting matrix to obtain an initial generated image; generating an all-zero image consistent with the to-be-processed SAR image in size, and generating a background image by using the all-zero image and the to-be-processed SAR image; and performing background filling processing on the initially generated image by using the background image to obtain a final fused SAR image. The resolution information and detail integrity of the to-be-processed SAR image are improved, and the calculation efficiency of sidelobe suppression is improved at the same time.

Description

Technical Field

[0001] The present invention relates to the field of radar image processing technology, and in particular to a high-resolution SAR sidelobe suppression method and device based on compressed sensing. Background Art

[0002] Synthetic Aperture Radar (SAR), as a high-resolution microwave imaging technology, has been widely used in the field of target detection and recognition due to its all-day, all-weather working characteristics and strong penetration ability. However, in the imaging process of SAR images, due to the limitations of the system's two-dimensional support domain in the frequency domain, the impulse response function shows the characteristics of a sinc function in both the azimuth and distance directions, which in turn produces excessively high sidelobe levels in the image, especially for large targets such as ships and buildings, which will form obvious "cross" spots. These spots not only interfere with the accurate extraction of target information, but also seriously reduce the accuracy of image interpretation, posing a challenge to subsequent target detection and recognition. Therefore, how to effectively suppress the sidelobes in SAR images has become a key issue in current SAR image preprocessing technology.

[0003] To solve the problem of SAR image sidelobe suppression, the Stereo Variable Area (SVA) algorithm is widely used in this field as a common nonlinear processing method. The SVA algorithm applies different weights to each sample in the SAR image. These weights are calculated based on the values ​​of the sample and its adjacent samples to distinguish between the main lobe and the side lobe. When there are strong targets or sinc function features in the image, SVA uses nonlinear operations to retain the main lobe and remove the side lobes; when the image contains multiple strong targets or multiple sinc functions, a linear filter is applied for processing. This method can effectively suppress the side lobes to a certain extent and improve the clarity of the image.

[0004] Although the SVA algorithm has achieved certain results in SAR image sidelobe suppression, it still has some shortcomings. First, while suppressing the sidelobes, the SVA algorithm cannot accurately restore the sample values ​​in the main lobe, which may lead to partial loss of target information in the image. Secondly, the SVA algorithm directly sets the sample values ​​on the sidelobes to zero and smoothes the image through median filtering. Although this approach reduces the impact of the sidelobes, it will inevitably reduce the resolution of the image and affect the detail performance of the image. In addition, although the existing linear windowing method can suppress the sidelobes to an extremely low level, it will also bring about a significant decrease in spatial resolution and is not suitable for the preprocessing requirements of high-resolution SAR images. Therefore, how to effectively suppress the sidelobes while maintaining the high resolution and detail integrity of the image is still an urgent problem to be solved in the current SAR image preprocessing technology. Summary of the invention

[0005] In order to solve the above problems existing in the prior art, the present invention provides a high-resolution SAR sidelobe suppression method and device based on compressed sensing.

[0006] The technical problem to be solved by the present invention is achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a high-resolution SAR sidelobe suppression method based on compressed sensing, comprising:

[0008] Acquire a SAR image to be processed, and determine a weighting function based on the SAR image to be processed and an SVA algorithm;

[0009] Generate filtered SAR images using weighting functions;

[0010] Acquire a measurement matrix, and use the filtered SAR image to clip the measurement matrix to obtain a measurement clipping matrix;

[0011] The main lobe value of the filtered SAR image is restored by using the inverse Moore-Penrose and measurement clipping matrix to obtain the initial generated image.

[0012] Generate an all-zero image with the same size as the SAR image to be processed, and generate a background image using the all-zero image and the SAR image to be processed;

[0013] The background image is used to perform background filling processing on the initial generated image to obtain the final fused SAR image.

[0014] Optionally, obtaining a SAR image to be processed, and determining a weighting function based on the SAR image to be processed and an SVA algorithm includes:

[0015] Acquire the SAR image to be processed;

[0016] Perform upsampling processing on the SAR image to be processed to obtain an upsampled SAR image;

[0017] The weighting function is generated using the upsampled SAR image and the SVA algorithm.

[0018] Optionally, the complex representation of the upsampled SAR image is:

[0019] g(m,n)=I(m,n)+iQ(m,n);

[0020] g(m,n) represents the complex number representation of the pixel in the mth row and nth column of the upsampled SAR image, I(m,n) represents the real part of g(m,n), Q(m,n) represents the imaginary part of g(m,n), i represents the imaginary unit, m≤M, n≤N, M represents the total number of elements of the column vector corresponding to the SAR image to be processed, and N represents the total number of elements of the row vector corresponding to the SAR image to be processed;

[0021] The weighting function includes: a real part weighting function and an imaginary part weighting function;

[0022] The real part weighting function is expressed as:

[0023]

[0024] Among them, w I (m,n) represents the real part weighting function of the pixel in the mth row and nth column of the upsampled SAR image, I(m,nc) represents the real part of the pixel in the mth row and ncth column, I(m,n+c) represents the real part of the pixel in the mth row and n+cth column, and c represents the offset;

[0025] The imaginary weighting function is expressed as:

[0026]

[0027] Among them, w Q (m,n) represents the imaginary part weighting function of the pixel in the mth row and nth column of the upsampled SAR image, Q(m,nc) represents the imaginary part of the pixel in the mth row and ncth column, Q(m,n+c) represents the imaginary part of the pixel in the mth row and n+cth column, and c represents the offset.

[0028] Optionally, generating a filtered SAR image using a weighting function includes:

[0029] Generate a real sparse value and an imaginary sparse value by using a weighting function;

[0030] A filtered SAR image is generated based on the real part sparse value and the imaginary part sparse value.

[0031] Optionally, the filtered SAR image is represented as:

[0032]

[0033] Among them, g sparse (m) represents the SAR image after the mth row of filtering, I sparse (m) represents the real sparse value of the upsampled SAR image in the mth row, Q sparse (m) represents the sparse value of the imaginary part of the upsampled SAR image in the mth row;

[0034]

[0035] w I (m) represents w I (m,n) corresponds to the entire row of data, w Q (m) represents w Q (m,n) corresponds to the entire row of data, w I (m,n) represents the real part weighting function of the pixel in the mth row and nth column of the upsampled SAR image, w Q (m,n) represents the imaginary weighting function of the pixel at the mth row and nth column of the upsampled SAR image.

[0036] Optionally, a measurement matrix is ​​obtained, and the measurement matrix is ​​clipped using the filtered SAR image to obtain a measurement clipping matrix, including:

[0037] Get a measurement matrix Ψ of size H×N;

[0038] Take the indexes of all non-zero elements in each row of the filtered SAR image to obtain an index set;

[0039] Clip the column vectors of the corresponding positions of the measurement matrix according to the index set, and form a measurement clipping matrix;

[0040] Where H represents the total number of elements in the column vector of the measurement matrix, and N represents the total number of elements in the row vector of the measurement matrix; where δ≤H≤N, δ represents the sparsity of g(m), g(m) represents the mth row vector of g(m,n), the total number of elements in the row vector of the measurement matrix is ​​equal to the total number of elements in the row vector of the SAR image to be processed, and the sparsity of g(m) is the number of all non-zero elements in g(m).

[0041] Optionally, generating an all-zero image having the same size as the SAR image to be processed, and generating a background image using the all-zero image and the SAR image to be processed, includes:

[0042] Generate an all-zero image with the same size as the SAR image to be processed;

[0043] The SAR image to be processed and the all-zero image are divided and processed based on a preset division rule, and a first divided image and a second divided image are obtained correspondingly;

[0044] Based on the first segmented image and the second segmented image, a background image is generated by adopting minimum modulus and cubic two-dimensional interpolation.

[0045] Optionally, based on the first segmented image and the second segmented image, a background image is generated by using a minimum modulus and cubic two-dimensional interpolation, including:

[0046] Calculate the modulus values ​​of all pixels of the first divided image in each divided area, and take the minimum modulus value of the pixels in each divided area as the element value of the second divided image at the center point of the divided area, to obtain the second divided processed image;

[0047] The second divided processed image is interpolated using cubic two-dimensional interpolation to obtain a background image.

[0048] Optionally, the initial generated image is subjected to background filling processing using the background image to obtain a final fused SAR image, including:

[0049] Calculate the pixel modulus value of each pixel of the initial generated image;

[0050] The background image is used to fill the pixels with zero pixel modulus value to obtain the final fused SAR image.

[0051] In a second aspect, the present invention provides a high-resolution SAR sidelobe suppression device based on compressed sensing, the high-resolution SAR sidelobe suppression device based on compressed sensing includes: an acquisition unit, a generation unit, a clipping unit, a main lobe recovery unit and a filling unit;

[0052] The acquisition unit is used to: acquire the SAR image to be processed, and determine the weighting function based on the SAR image to be processed and the SVA algorithm;

[0053] The generating unit is used for: generating a filtered SAR image by using a weighting function;

[0054] The clipping unit is used to: obtain a measurement matrix, and clip the measurement matrix using the filtered SAR image to obtain a measurement clipping matrix;

[0055] The main lobe recovery unit is used to: perform main lobe value recovery processing on the filtered SAR image using the Moore-Penrose inverse and the measurement clipping matrix to obtain an initial generated image;

[0056] The generating unit is further used to: generate an all-zero image having the same size as the SAR image to be processed, and generate a background image using the all-zero image and the SAR image to be processed;

[0057] The filling unit is used to perform background filling processing on the initial generated image using the background image to obtain the final fused SAR image.

[0058] The present invention provides a high-resolution SAR sidelobe suppression method and device based on compressed sensing. The high-resolution SAR sidelobe suppression method based on compressed sensing includes: obtaining a SAR image to be processed, determining a weighting function based on the SAR image to be processed and an SVA algorithm; generating a filtered SAR image using the weighting function; obtaining a measurement matrix, and using the filtered SAR image to perform a clipping process on the measurement matrix to obtain a measurement clipping matrix; using the Moore-Penrose inverse and the measurement clipping matrix to perform a main lobe value recovery process on the filtered SAR image to obtain an initial generated image; generating an all-zero image of the same size as the SAR image to be processed, and using the all-zero image and the SAR image to be processed to generate a background image; using the background image to perform a background filling process on the initial generated image to obtain a final fused SAR image. In the present invention, the SVA algorithm is used to perform sparse processing on the SAR image to be processed, which can accurately distinguish the main lobe and the side lobe in the SAR image to be processed, and obtain the filtered SAR image; in addition, the background image filling can retain the contour information of the image better than the ordinary weighted average, thereby improving the resolution information and detail integrity of the SAR image to be processed; and the Moore-Penrose inverse and measurement clipping matrix are used to restore the filtered SAR image, which has higher computational efficiency than the traditional CS algorithm of l0 and l1 norms.

[0059] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of a flow chart of a high-resolution SAR sidelobe suppression method based on compressed sensing provided in an embodiment of the present invention;

[0061] Figure 2 A complete execution block diagram of a high-resolution SAR sidelobe suppression method based on compressed sensing provided by another embodiment of the present invention;

[0062] Figure 3 A schematic diagram of a SAR image to be processed is shown exemplarily;

[0063] Figure 4 The schematic diagram of the result after the existing SVA algorithm is used to perform sidelobe suppression on the SAR image to be processed is shown as an example;

[0064] Figure 5 The schematic diagram of the result after the existing CS algorithm is used to perform sidelobe suppression on the SAR image to be processed is shown as an example;

[0065] Figure 6 The schematic diagram of the result after the sidelobe suppression is performed on the SAR image to be processed by the method of the present invention is exemplarily shown;

[0066] Figure 7 A schematic structural diagram of a high-resolution SAR sidelobe suppression device based on compressed sensing provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] The present invention is further described in detail below with reference to specific embodiments, but the embodiments of the present invention are not limited thereto.

[0068] In order to effectively suppress side lobes while maintaining high resolution and detail integrity of an image, an embodiment of the present invention provides a high-resolution SAR side lobe suppression method based on compressed sensing. Figure 1 A schematic diagram of a flow chart of a high-resolution SAR sidelobe suppression method based on compressed sensing provided by an embodiment of the present invention. Figure 1 As shown, the method includes:

[0069] S101, obtaining a SAR image to be processed, and determining a weighting function based on the SAR image to be processed and an SVA algorithm.

[0070] Specifically, the SAR image G to be processed is a SAR image with a size of M×N and including sidelobe effects, M represents the total number of elements of the column vector corresponding to the SAR image to be processed, and N represents the total number of elements of the row vector corresponding to the SAR image to be processed.

[0071] Optionally, S101 may specifically include:

[0072] Acquire the SAR image to be processed;

[0073] Perform upsampling processing on the SAR image to be processed to obtain an upsampled SAR image;

[0074] The weighting function is generated using the upsampled SAR image and the SVA algorithm.

[0075] Optionally, the complex representation of the upsampled SAR image is:

[0076] g(m,n)=I(m,n)+iQ(m,n);

[0077] g(m,n) represents the complex number representation of the pixel in the mth row and nth column of the upsampled SAR image, I(m,n) represents the real part of g(m,n), Q(m,n) represents the imaginary part of g(m,n), i represents the imaginary unit, m≤M, n≤N, M represents the total number of elements of the column vector corresponding to the SAR image to be processed, and N represents the total number of elements of the row vector corresponding to the SAR image to be processed;

[0078] The weighting function includes: a real part weighting function and an imaginary part weighting function;

[0079] The real part weighting function is expressed as:

[0080]

[0081] Among them, w I (m,n) represents the real part weighting function of the pixel in the mth row and nth column of the upsampled SAR image, I(m,nc) represents the real part of the pixel in the mth row and ncth column, I(m,n+c) represents the real part of the pixel in the mth row and n+cth column, and c represents the offset;

[0082] The imaginary weighting function is expressed as:

[0083]

[0084] Among them, w Q (m,n) represents the imaginary part weighting function of the pixel in the mth row and nth column of the upsampled SAR image, Q(m,nc) represents the imaginary part of the pixel in the mth row and ncth column, Q(m,n+c) represents the imaginary part of the pixel in the mth row and n+cth column, and c represents the offset.

[0085] Furthermore, it is necessary to satisfy nc≤N, n+c≤N.

[0086] S102: Generate a filtered SAR image using a weighting function.

[0087] Optionally, S102 may specifically include:

[0088] Generate a real sparse value and an imaginary sparse value by using a weighting function;

[0089] A filtered SAR image is generated based on the real part sparse value and the imaginary part sparse value.

[0090] Optionally, the filtered SAR image is represented as:

[0091]

[0092] Among them, g sparse (m) represents the SAR image after the mth row of filtering, I sparse (m) represents the real sparse value of the upsampled SAR image in the mth row, Q sparse (m) represents the sparse value of the imaginary part of the upsampled SAR image in the mth row;

[0093]

[0094]

[0095] w I (m) represents w I (m,n) corresponds to the entire row of data, w Q(m) represents w Q (m,n) corresponds to the entire row of data, w I (m,n) represents the real part weighting function of the pixel in the mth row and nth column of the upsampled SAR image, w Q (m,n) represents the imaginary weighting function of the pixel at the mth row and nth column of the upsampled SAR image.

[0096] S103, obtaining a measurement matrix, and using the filtered SAR image to perform clipping processing on the measurement matrix to obtain a measurement clipping matrix.

[0097] Optionally, S103 may specifically include:

[0098] Get a measurement matrix Ψ of size H×N;

[0099] Take the indexes of all non-zero elements in each row of the filtered SAR image to obtain an index set;

[0100] Clip the column vectors of the corresponding positions of the measurement matrix according to the index set, and form a measurement clipping matrix;

[0101] Where H represents the total number of elements in the column vector of the measurement matrix, and N represents the total number of elements in the row vector of the measurement matrix; where δ≤H≤N, δ represents the sparsity of g(m), g(m) represents the mth row vector of g(m,n), the total number of elements in the row vector of the measurement matrix is ​​equal to the total number of elements in the row vector of the SAR image to be processed, and the sparsity of g(m) is the number of all non-zero elements in g(m).

[0102] S104, using the Moore-Penrose inverse and the measurement clipping matrix to perform main lobe value recovery processing on the filtered SAR image to obtain an initial generated image.

[0103] Specifically, the initial generated image The mth row vector of It can be expressed as:

[0104]

[0105] Among them, p represents the measurement clipping matrix, g cs (m) represents the image after g(m) compression;

[0106] g cs (m) = Ψ·g(m);

[0107] Among them, Ψ represents the measurement matrix, g(m) represents the mth row vector of g(m,n), consists of non-zero samples.

[0108] S105 , generating an all-zero image having the same size as the SAR image to be processed, and generating a background image using the all-zero image and the SAR image to be processed.

[0109] Optionally, S105 may specifically include:

[0110] Generate an all-zero image with the same size as the SAR image to be processed;

[0111] The SAR image to be processed and the all-zero image are divided and processed based on a preset division rule, and a first divided image and a second divided image are obtained correspondingly;

[0112] Based on the first segmented image and the second segmented image, a background image is generated by adopting minimum modulus and cubic two-dimensional interpolation.

[0113] In this processing process: First, initialize the image G with the same size as the SAR image G to be processed and with all pixel values ​​zero bg , and the SAR image G to be processed and the all-zero image G bg The first divided images are uniformly divided into I×J image regions of size X×Y, and the first divided images are obtained respectively. and the second partitioned image I represents the total number of horizontally divided regions, J represents the total number of vertically divided regions, X represents the total number of column elements in a column of divided regions, and Y represents the total number of row elements in a row of divided regions. and the second partitioned image It is expressed as follows:

[0114]

[0115] Among them G ij express The image area of ​​the i-th row and j-th column, G bg_ij express The image area of ​​the i-th row and j-th column, I≥2, J≥2.

[0116] Optionally, based on the first segmented image and the second segmented image, a background image is generated by using a minimum modulus and cubic two-dimensional interpolation, including:

[0117] Calculate the modulus values ​​of all pixels of the first divided image in each divided area, and take the minimum modulus value of the pixels in each divided area as the element value of the second divided image at the center point of the divided area, to obtain the second divided processed image;

[0118] The second divided processed image is interpolated using cubic two-dimensional interpolation to obtain a background image.

[0119] Specifically, the process of obtaining the background image is as follows: Calculate the image area Gij The modulus value V of each pixel in ij (x,y), and select G ij The minimum modulus V of the region min_ij As the image area G bg_ij The value of the center point, where x ≤ X, y ≤ Y.

[0120]

[0121] I ij (x,y),Q ij (x, y) represent the image region G ij The real and imaginary parts of the pixel at the xth row and yth column in .

[0122] The center point value is V in the azimuth and distance directions respectively. min_ij The image area G bg_ij Perform cubic two-dimensional interpolation processing to obtain the background image G' bg , background image G' bg It can be expressed as:

[0123]

[0124] Among them, G' bg_ij Represents the background image G' bg The image area of ​​the i-th row and j-th column.

[0125] S106: Perform background filling processing on the initial generated image using the background image to obtain a final fused SAR image.

[0126] Optionally, S106 may specifically include:

[0127] Calculate the pixel modulus value of each pixel of the initial generated image;

[0128] The background image is used to fill the pixels with zero pixel modulus value to obtain the final fused SAR image.

[0129] Specifically, through the background image G' bg The value g' of the mth row and nth column bg (m,n) pair of initial generated graphs

[0130] picture The pixel points with pixel modulus value of zero in the image are filled to obtain the final fused SAR image G out :

[0131]

[0132] Among them, g out (m,n) represents the final fused SAR image G outThe pixel value of the mth row and nth column, is the initial generated image The value of the mth row and nth column, g' bg (m,n) is the background image G' bg The value of the mth row and nth column of .

[0133] The embodiment of the present invention provides a high-resolution SAR sidelobe suppression method based on compressed sensing. The SVA algorithm is used to perform sparse processing on the SAR image to be processed, and the main lobe and side lobe in the SAR image to be processed can be accurately distinguished to obtain a filtered SAR image. In addition, background image filling can better retain the contour information of the image than ordinary weighted averaging, thereby improving the resolution information and detail integrity of the SAR image to be processed. The Moore-Penrose inverse and measurement clipping matrix are used to restore the filtered SAR image, which has higher calculation efficiency than the traditional CS algorithm of l0 and l1 norms.

[0134] Figure 2 A complete execution block diagram of a high-resolution SAR sidelobe suppression method based on compressed sensing provided by another embodiment of the present invention. Figure 2 As shown in the figure, firstly, a SAR image is acquired, and then the SAR image is preprocessed, and a weighting function of the real part and the imaginary part of each pixel is defined according to the preprocessing result; then, a filtered SAR image is generated according to the weighting function, and the measurement matrix is ​​clipped by using the filtered SAR image to obtain a measurement clipping matrix, and then the main lobe value of the SAR image is restored by using the measurement clipping matrix and Moore-Penrose inverse to obtain a SAR image with restored main lobe value, and finally, a background image is constructed, and the background of the SAR image with restored main lobe value is filled in combination with the background image to obtain the final fused SAR image.

[0135] In summary, the present invention has the following advantages compared with the prior art:

[0136] 1. The high-resolution SAR sidelobe suppression method based on compressed sensing of the present invention is adopted. When the weighted function of the real part and the weighted function of the imaginary part of each pixel are both greater than their upper limit and less than their lower limit, the corresponding pixel is set to zero, thereby realizing filtering of the real part and the imaginary part of each pixel of the upsampled SAR image, calculating the position of the main lobe of the image, and then clipping the measurement matrix according to the main lobe position, restoring the main lobe of the compressed image using the Moore-Penrose inverse, and then filling the zero value with the background image. The method fully combines the SVA and Moore-Penrose inverse algorithms, realizes the accurate restoration of the main lobe of the SAR image to be processed, can retain the contour information of the image better than the ordinary weighted average, and improves the sidelobe suppression capability of the SAR image to be processed.

[0137] 2. Using the Moore-Penrose inverse and the measured cropping matrix to restore the processed SAR image has higher computational efficiency than the traditional l0 and l1 norm CS algorithm.

[0138] In order to illustrate the effectiveness of the high-resolution SAR sidelobe suppression method based on compressed sensing provided by the present invention, a simulation experiment is also carried out as follows:

[0139] It should be noted that the indicators used to measure the degree of sidelobe suppression in the field of SAR image sidelobe suppression are the brightness of the luminous cross on the target in the SAR image and the resolution of the image after sidelobe suppression. Among them, the darker the luminous cross around the target after sidelobe suppression, the better the sidelobe suppression effect and the higher the accuracy of target interpretation; the closer the image clarity after sidelobe suppression is to the original image, the less image information is lost by sidelobe suppression, and the higher the image resolution after sidelobe suppression. Figure 3 The schematic diagram of the SAR image to be processed is shown as an example. Figure 3 As shown in Figure 1, there is an object similar to a solar panel in the center of the SAR image to be processed. Due to the strong side lobes, the points on the panel are almost connected, and the surrounding metal structures are also somewhat degraded. Figure 4 The schematic diagram exemplarily shows the result of sidelobe suppression of the SAR image to be processed using the existing SVA algorithm. Figure 5 The schematic diagram exemplarily shows the result of sidelobe suppression of the SAR image to be processed using the existing CS algorithm. Figure 6 The schematic diagram exemplarily shows the result of sidelobe suppression of the SAR image to be processed by the method of the present invention. Figure 4 , Figure 5 and Figure 6 It can be seen from the simulation results that Figure 4 The SVA algorithm can suppress the sidelobe effect of the target on the SAR image to be processed to a certain extent, but Figure 4 There are still intermittent bright spots in the video, and the image quality is poor. Figure 5 The use of the CS algorithm reduces the clarity of the SAR image to be processed, and a large number of noise points are scattered in the image. Figure 6 The experimental results based on the method of the present invention make the point features on the plate very obvious, the surrounding structures are restored, and the background is almost unchanged. In contrast, because the method of the present invention fully combines the Moore-Penrose inverse and SVA prior mechanisms, it achieves better results in suppressing the side lobes, accurately retains the information of the main lobe, and retains the background information of the SAR image to be processed after adding the background image filling, making Figure 6 The contour information in the image is very obvious, and the image resolution is not significantly reduced, which effectively improves the quality of the SAR image to be processed.

[0140] The method provided in the embodiment of the present invention can be applied to an electronic device. Specifically, the electronic device can be: a desktop computer, a portable computer, an intelligent mobile terminal, a server, etc., which is not limited in the embodiment of the present invention.

[0141] Based on the same inventive concept, an embodiment of the present invention further provides a high-resolution SAR sidelobe suppression device based on compressed sensing. Figure 7 A schematic diagram of the structure of a high-resolution SAR sidelobe suppression device based on compressed sensing provided by an embodiment of the present invention. Figure 7 As shown, it includes: an acquisition unit 701, a generation unit 702, a cropping unit 703, a main lobe recovery unit 704 and a filling unit 705;

[0142] The acquisition unit 701 is used to: acquire the SAR image to be processed, and determine the weighting function based on the SAR image to be processed and the SVA algorithm;

[0143] The generating unit 702 is used to: generate a filtered SAR image using a weighting function;

[0144] The clipping unit 703 is used to: obtain a measurement matrix, and clip the measurement matrix using the filtered SAR image to obtain a measurement clipping matrix;

[0145] The main lobe recovery unit 704 is used to: perform main lobe value recovery processing on the filtered SAR image using the Moore-Penrose inverse and measurement clipping matrix to obtain an initial generated image;

[0146] The generating unit 702 is further used to: generate an all-zero image having the same size as the SAR image to be processed, and generate a background image using the all-zero image and the SAR image to be processed;

[0147] The filling unit 705 is used to perform background filling processing on the initially generated image using the background image to obtain a final fused SAR image.

[0148] It should be noted that the terms "first", "second", etc. are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention 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 invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention.

[0149] 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 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 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.

[0150] Although the present invention is described herein in conjunction with various embodiments, in the process of implementing the claimed invention, those skilled in the art can understand and implement other changes of the above disclosed embodiments by viewing the drawings and the disclosed content. In the description of the present invention, the term "comprising" does not exclude other components or steps, "one" or "an" does not exclude multiple situations, and the meaning of "multiple" is two or more, unless otherwise clearly and specifically limited. In addition, certain measures are recorded in different embodiments, but this does not mean that these measures cannot be combined to produce good results.

[0151] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the scope of protection of the present invention.

Claims

1. A high-resolution SAR sidelobe suppression method based on compressed sensing, characterized in that: include: Acquire a SAR image to be processed, and determine a weighting function based on the SAR image to be processed and an SVA algorithm; generating a filtered SAR image using the weighting function; Acquire a measurement matrix, and use the filtered SAR image to perform clipping processing on the measurement matrix to obtain a measurement clipping matrix; Using the Moore-Penrose inverse and the measurement clipping matrix to perform main lobe value recovery processing on the filtered SAR image to obtain an initial generated image; Generate an all-zero image with the same size as the SAR image to be processed, and generate a background image using the all-zero image and the SAR image to be processed; The background image is used to perform background filling processing on the initially generated image to obtain a final fused SAR image.

2. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 1 is characterized in that: The step of acquiring a SAR image to be processed and determining a weighting function based on the SAR image to be processed and an SVA algorithm includes: Acquire the SAR image to be processed; Performing upsampling processing on the SAR image to be processed to obtain an upsampled SAR image; The weighting function is generated using the upsampled SAR image and an SVA algorithm.

3. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 2 is characterized in that: The complex representation of the upsampled SAR image is: g(m,n)=I(m,n)+iQ(m,n); g(m,n) represents the complex number representation of the pixel point at the mth row and nth column of the upsampled SAR image, I(m,n) represents the real part of g(m,n), Q(m,n) represents the imaginary part of g(m,n), i represents the imaginary unit, m≤M, n≤N, M represents the total number of elements of the column vector corresponding to the SAR image to be processed, and N represents the total number of elements of the row vector corresponding to the SAR image to be processed; The weighting function includes: a real part weighting function and an imaginary part weighting function; The real part weighting function is expressed as: Among them, w I (m,n) represents the real part weighting function of the pixel point in the mth row and nth column of the upsampled SAR image, I(m,nc) represents the real part of the pixel point in the mth row and ncth column, I(m,n+c) represents the real part of the pixel point in the mth row and n+cth column, and c represents the offset; The imaginary part weighting function is expressed as: Among them, w Q (m,n) represents the imaginary part weighting function of the pixel point in the mth row and nth column of the upsampled SAR image, Q(m,nc) represents the imaginary part of the pixel point in the mth row and ncth column, Q(m,n+c) represents the imaginary part of the pixel point in the mth row and n+cth column, and c represents the offset.

4. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 1 is characterized in that: The step of generating a filtered SAR image using the weighting function comprises: Generating a real part sparse value and an imaginary part sparse value by using the weighting function; The filtered SAR image is generated based on the real part sparse value and the imaginary part sparse value.

5. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 4 is characterized in that: The filtered SAR image is expressed as: Among them, g sparse (m) represents the SAR image after the mth row of filtering, I sparse (m) represents the real sparse value of the upsampled SAR image in the mth row, Q sparse (m) represents the sparse value of the imaginary part of the upsampled SAR image in the mth row; w I (m) represents w I (m,n) corresponds to the entire row of data, w Q (m) represents w Q (m,n) corresponds to the entire row of data, w I (m,n) represents the real part weighting function of the pixel point in the mth row and nth column of the upsampled SAR image, w Q (m,n) represents the imaginary part weighting function of the pixel at the mth row and nth column of the upsampled SAR image.

6. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 1 is characterized in that: The obtaining of the measurement matrix and using the filtered SAR image to perform clipping processing on the measurement matrix to obtain a measurement clipping matrix includes: Obtain the measurement matrix Ψ of size H×N; Taking the indexes of all non-zero elements in each row of the filtered SAR image to obtain an index set; Clipping the column vectors at corresponding positions of the measurement matrix according to the index set, and forming the measurement clipping matrix; Wherein, H represents the total number of elements of the column vector of the measurement matrix, and N represents the total number of elements of the row vector of the measurement matrix; wherein, δ≤H≤N, δ represents the sparsity of g(m), g(m) represents the mth row vector where g(m,n) is located, the total number of elements of the row vector of the measurement matrix is ​​equal to the total number of elements of the row vector of the SAR image to be processed, and the sparsity of g(m) is the number of all non-zero elements in g(m).

7. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 1 is characterized in that: The step of generating an all-zero image having the same size as the SAR image to be processed, and generating a background image using the all-zero image and the SAR image to be processed includes: Generate the all-zero image that is consistent with the size of the SAR image to be processed; The SAR image to be processed and the all-zero image are divided and processed based on a preset division rule to obtain a first divided image and a second divided image; Based on the first segmented image and the second segmented image, the background image is generated by adopting minimum modulus and cubic two-dimensional interpolation.

8. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 7 is characterized in that: The step of generating the background image based on the first segmented image and the second segmented image by using the minimum modulus and the cubic two-dimensional interpolation comprises: Calculating the modulus values ​​of all pixels of the first divided image in each divided area, and taking the minimum modulus value of the pixels in each divided area as the element value of the second divided image at the center point of the divided area, to obtain a second divided processed image; The second divided processed image is interpolated using the cubic two-dimensional interpolation to obtain the background image.

9. The high-resolution SAR sidelobe suppression method based on compressed sensing according to claim 1, characterized in that: The using the background image to perform background filling processing on the initially generated image to obtain a final fused SAR image includes: Calculate the pixel modulus value of each pixel point of the initially generated image; The pixel points whose pixel modulus value is zero are filled with the background image to obtain the final fused SAR image.

10. A high-resolution SAR sidelobe suppression device based on compressed sensing, characterized in that: The high-resolution SAR sidelobe suppression device based on compressed sensing includes: an acquisition unit, a generation unit, a clipping unit, a main lobe recovery unit and a filling unit; The acquisition unit is used to: acquire a SAR image to be processed, and determine a weighting function based on the SAR image to be processed and an SVA algorithm; The generating unit is used to: generate a filtered SAR image using the weighting function; The clipping unit is used to: obtain a measurement matrix, and use the filtered SAR image to clip the measurement matrix to obtain a measurement clipping matrix; The main lobe recovery unit is used to: use the Moore-Penrose inverse and the measurement clipping matrix to perform main lobe value recovery processing on the filtered SAR image to obtain an initial generated image; The generating unit is further used to: generate an all-zero image having the same size as the SAR image to be processed, and generate a background image using the all-zero image and the SAR image to be processed; The filling unit is used to: perform background filling processing on the initially generated image using the background image to obtain a final fused SAR image.

Images