A SAR image sidelobe suppression method, device, electronic device and storage medium

Abstract

The embodiments of the present invention relate to the field of image processing technology, and in particular to a sidelobe suppression method, device, electronic device and storage medium for SAR images. The method comprises: calculating a two-dimensional PSF estimation matrix of a SAR image to be processed; solving a deconvolution function based on the two-dimensional PSF matrix and relevant parameters of the SAR image to be processed, and outputting a deconvolution result; wherein the deconvolution result can be restored to a SAR image after sidelobe suppression, and the relevant parameters include the number of sampling points in the azimuth direction, the number of sampling points in the range direction, the phase error coefficient, the blur kernel size, the penalty parameter, the equilibrium constant and the maximum number of iterations of the SAR image to be processed, the azimuth direction is the direction of the radar flight track, and the range direction is the direction along which the radar wave is emitted. The sidelobe suppression method for SAR images provided in the present application has good sidelobe suppression performance, and can effectively filter out noise and improve image quality.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method, device, electronic equipment and storage medium for sidelobe suppression of SAR images. Background Art

[0002] The sidelobe suppression problem is an important issue in traditional synthetic aperture radar image processing. Synthetic aperture radar obtains two-dimensional high-resolution images through range pulse compression technology and azimuth synthetic aperture technology. However, since the spectrum support domain of the echo signal received by the SAR system is finitely distributed, the point spread function (PSF) of the SAR image is a Sinc function in both range and azimuth, and there are non-negligible sidelobes.

[0003] In SAR images, strong targets, such as buildings and metal objects, not only have strong main lobe energy, but also have high side lobe energy, which can easily exceed the main lobe energy of surrounding weak targets. Weak targets are hidden under the side lobes of strong targets and cannot be identified. Therefore, it is necessary to suppress the side lobes of SAR images to improve the image quality and visual readability of SAR images, so as to facilitate subsequent image interpretation.

[0004] In the prior art, the commonly used SAR image sidelobe suppression methods mainly include frequency domain weighting method, apodization filtering method and spatially variant apodization (SVA). However, these methods have limited performance in SAR image sidelobe suppression, poor sidelobe suppression effect, and the quality of processed images still has a lot of room for improvement. Therefore, a new SAR image sidelobe suppression method is urgently needed to solve the above problems. Summary of the invention

[0005] In order to effectively suppress the sidelobe noise of a SAR image, an embodiment of the present invention provides a sidelobe suppression method, device, electronic device and storage medium for a SAR image.

[0006] In a first aspect, an embodiment of the present invention provides a sidelobe suppression method for a SAR image, comprising:

[0007] Calculate the two-dimensional PSF estimation matrix of the SAR image to be processed;

[0008] Based on the two-dimensional PSF estimation matrix and the relevant parameters of the SAR image to be processed, a deconvolution function is solved and a deconvolution result is output; wherein the deconvolution result can be restored to a SAR image after sidelobe suppression, and the relevant parameters include the number of sampling points in the azimuth direction, the number of sampling points in the range direction, the phase error coefficient, the blur kernel size, the penalty parameter, the equilibrium constant and the maximum number of iterations of the SAR image to be processed, the azimuth direction is the direction of the radar flight track, and the range direction is the direction along which the radar wave is emitted.

[0009] In a possible design, calculating a two-dimensional PSF estimation matrix of a SAR image to be processed includes:

[0010] Acquiring a defocused image of the SAR image to be processed;

[0011] Calculating a first factor of the SAR image to be processed based on the SAR image to be processed and a defocused image of the SAR image to be processed;

[0012] Calculating a second factor of the SAR image to be processed based on the SAR image to be processed and a defocused image of the SAR image to be processed;

[0013] Calculating an auxiliary factor of the SAR image to be processed based on the first factor and the second factor;

[0014] Perform a two-dimensional inverse Fourier transform on the auxiliary factors to obtain a two-dimensional PSF estimation matrix of the SAR image to be processed.

[0015] In a possible design, acquiring the defocused image of the SAR image to be processed includes:

[0016] Read in the SAR image to be processed and related parameters of the SAR image to be processed; wherein the related parameters include the number of azimuth sampling points N of the SAR image to be processed a , the number of sampling points in the distance direction N r 、Phase error coefficient [Ep a , Ep r ], blur kernel size [h r ,h a ], penalty parameter μ, equilibrium constant λ and maximum number of iterations Iter max ;

[0017] Performing a two-dimensional Fourier transform on the SAR image to be processed to obtain a first frequency domain image;

[0018] Adding a phase error to the first frequency domain image to obtain a second frequency domain image containing the phase error;

[0019] Perform a two-dimensional inverse Fourier transform on the second frequency domain image to obtain a defocused image of the SAR image to be processed.

[0020] In a possible design, calculating the first factor of the SAR image to be processed based on the SAR image to be processed and the defocused image of the SAR image to be processed includes:

[0021] Performing range Fourier transform on the SAR image to be processed and the defocused image of the SAR image to be processed respectively to obtain a first image and a second image; wherein each row vector in the first image can be expressed as Each row vector in the second image can be expressed as Where k = 1, 2, ..., N a , l = 1, 2, ..., N r ;

[0022] Calculate the auxiliary factor of each row vector in the first image separately The first factor A(k, l) of the SAR image to be processed is obtained; wherein,

[0023] In a possible design, calculating the second factor of the SAR image to be processed based on the SAR image to be processed and the defocused image of the SAR image to be processed includes:

[0024] Performing azimuth Fourier transform on the SAR image to be processed and the defocused image of the SAR image to be processed respectively to obtain a third image and a fourth image; wherein each row vector in the third image can be expressed as Each row vector in the fourth image can be expressed as Where k = 1, 2, ..., N a , l = 1, 2, ..., N r ;

[0025] Calculate the auxiliary factor of each row vector in the third image respectively The second factor B(k, l) of the SAR image to be processed is obtained; wherein,

[0026] In a possible design, calculating the auxiliary factor of the to-be-processed SAR image based on the first factor and the second factor includes:

[0027] The first factor A(k, l) and the second factor B(k, l) are combined to obtain a first equation: A(k, l)a(k)-B(k, l)b(l)=0; wherein,

[0028] The first equation is integrated into the second equation: Γ·t=0; where,

[0029]

[0030] t=(a(0) … a(N a -1) b(0) … b(N r -1)) T ;

[0031] The singular value decomposition method is used to solve the t, calculate the a(k) and the b(l), and obtain the auxiliary factor of the SAR image to be processed. Where k = 1, 2, ..., N a , l = 1, 2, ..., N r -1;

[0032] In a possible design, performing a two-dimensional inverse Fourier transform on the auxiliary factors of the SAR image to be processed to obtain a two-dimensional PSF estimation matrix of the SAR image to be processed includes:

[0033] Regarding the Perform a two-dimensional inverse Fourier transform to obtain a two-dimensional PSF estimation matrix T of the SAR image to be processed; wherein, = 1, 2, ..., N a , l = 1, 2, ..., N r .

[0034] In a possible design, solving a deconvolution function based on the two-dimensional PSF matrix and relevant parameters of the SAR image to be processed and outputting a deconvolution result includes:

[0035] Based on the number of azimuth sampling points N of the SAR image to be processed a And the number of sampling points N in the distance direction r , generate image Y; wherein the size of the image Y is (N a +h a )×(N r +h r ), and the SAR image to be processed is located at the center of the image Y, and in the matrix corresponding to the image Y, except for the pixel at the location of the SAR image to be processed, the other pixels are 0;

[0036] Construct the deconvolution function: y = Hx + n; where, is a one-dimensional vector generated by straightening the image Y, x is the deconvolution result to be calculated, n is independent and identically distributed Gaussian white noise, H is a vector with a size of N a N r ×N a N r A block circulant matrix H is calculated by the two-dimensional PSF matrix T;

[0037] make The deconvolution function is transformed into the following third set of equations:

[0038]

[0039]

[0040] Where N = N a ×N r , m=N a ×h r +N r ×h a +h a ×h r , is a one-dimensional vector generated by straightening the SAR image to be processed, z=zeros(m, 1) is a zero vector, (a) j represents the jth element of vector a, D h and D v Represents the first-order derivatives of the SAR image to be processed in azimuth and range directions, λ>0;

[0041] The third equation is iteratively calculated using the fast deconvolution method of the alternating direction multiplier method until the maximum number of iterations Iter is reached. max , output the deconvolution result x.

[0042] In a second aspect, an embodiment of the present invention further provides a sidelobe suppression device for a SAR image, comprising:

[0043] A calculation module, used for calculating a two-dimensional PSF estimation matrix of a SAR image to be processed;

[0044] A solution module is used to solve the deconvolution function based on the two-dimensional PSF estimation matrix and the relevant parameters of the SAR image to be processed, and output the deconvolution result; wherein the deconvolution result can be restored to the SAR image after sidelobe suppression, and the relevant parameters include the number of sampling points in the azimuth direction, the number of sampling points in the range direction, the phase error coefficient, the blur kernel size, the penalty parameter, the equilibrium constant and the maximum number of iterations of the SAR image to be processed, the azimuth direction is the direction of the radar flight track, and the range direction is the direction along the radar wave emission.

[0045] In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in any embodiment of this specification is implemented.

[0046] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, enables the computer to execute the method described in any embodiment of this specification.

[0047] The embodiment of the present invention provides a sidelobe suppression method, device, electronic device and storage medium for SAR images. First, a two-dimensional PSF estimation matrix of the SAR image to be processed is calculated; then, based on the two-dimensional PSF matrix and the relevant parameters of the SAR image to be processed, a deconvolution function is solved and a deconvolution result is output; finally, the deconvolution result is restored to a SAR image after sidelobe suppression. The present application suppresses the sidelobes in the SAR image by the deconvolution method, which can effectively filter out noise and improve image quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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.

[0049] Figure 1 This is a flow chart of a SAR image sidelobe suppression method provided by one embodiment of the present invention;

[0050] Figure 2 is a SAR image containing side lobes and Gaussian white noise provided by an embodiment of the present invention;

[0051] Figure 3 An embodiment of the present invention provides Figure 2 The two-dimensional PSF estimation matrix result;

[0052] Figure 4 This is an embodiment of the present invention, which uses this method to Figure 2 SAR image after sidelobe suppression;

[0053] Figure 5 An embodiment of the present invention provides Figure 2 Sidelobe suppression result profile of the target point 5 in the range direction;

[0054] Figure 6 An embodiment of the present invention provides Figure 2Sidelobe suppression result profile of the target point 5 in azimuth;

[0055] Figure 7 is a hardware architecture diagram of an electronic device provided by an embodiment of the present invention;

[0056] Figure 8 The present invention is a structural diagram of a SAR image sidelobe suppression device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0058] As mentioned above, the commonly used SAR image sidelobe suppression methods currently include frequency domain weighting method, apodization filtering method and spatially variant apodization (SVA). Among them, the frequency domain weighting method is a windowing method. By applying different weighting functions to the two-dimensional spectrum of the image in the frequency domain, the discontinuity at the edge of the spectrum is weakened, thereby reducing the sidelobe energy in the SAR image. The frequency domain weighting method is simple to implement and fast to process. It is widely used in engineering applications. However, the frequency domain weighting method will reduce or lose some frequency energy, resulting in main lobe broadening. The sidelobe suppression performance of this method is at the expense of resolution loss. The apodization filtering method is a nonlinear processing method. This method first generates a SAR image with low sidelobe and main lobe broadening after processing using the frequency domain weighting method, and then compares the original image and the weighted image. At each spatial position, the image with the smallest amplitude of the complex value is selected from the two images as the final result. The performance of the apodization filtering method is better than that of the frequency domain weighting method. It can simultaneously reduce the sidelobe and maintain the main lobe width, but its sidelobe suppression performance is limited. The spatial apodization method is an adaptive nonlinear processing method. This method adaptively selects a cosine weighting function suitable for each pixel in the spatial position to reduce the sidelobe energy. The spatial apodization method has the best overall sidelobe suppression effect. However, due to the existence of system distortion and phase noise, the PSF of the SAR image is no longer a regular two-dimensional Sinc function, resulting in performance loss of the spatial apodization method. It can be seen that the existing methods have limited sidelobe suppression performance for SAR images, and the sidelobe suppression effect is not good. There is still much room for improvement in the quality of the processed images.

[0059] In order to solve the above technical problems, it is possible to consider performing sidelobe suppression on SAR images by deconvolution, thereby improving the image quality of SAR images.

[0060] Please refer to Figure 1 The embodiment of the present invention provides a sidelobe suppression method for a SAR image, the method comprising:

[0061] Step 100: Calculate the two-dimensional PSF estimation matrix of the SAR image to be processed;

[0062] Step 102: Based on the two-dimensional PSF estimation matrix and the relevant parameters of the SAR image to be processed, solve the deconvolution function and output the deconvolution result; wherein the deconvolution result can be restored to the SAR image after sidelobe suppression, and the relevant parameters include the number of sampling points in the azimuth direction, the number of sampling points in the range direction, the phase error coefficient, the blur kernel size, the penalty parameter, the equilibrium constant and the maximum number of iterations of the SAR image to be processed, the azimuth direction is the direction of the radar flight track, and the range direction is the direction along the radar wave emission.

[0063] In the embodiment of the present invention, the two-dimensional PSF estimation matrix of the SAR image to be processed is calculated; then based on the two-dimensional PSF matrix and the relevant parameters of the SAR image to be processed, the deconvolution function is solved and the deconvolution result is output; finally, the deconvolution result is restored to the SAR image after sidelobe suppression. The SAR image sidelobe suppression method provided by the present invention can effectively suppress the sidelobe noise in the SAR image and improve the image quality.

[0064] Described below Figure 1 How the various steps are performed.

[0065] In some embodiments, step 100 includes:

[0066] Step A1: obtaining a defocused image of the SAR image to be processed;

[0067] Step A2: calculating a first factor of the SAR image to be processed based on the SAR image to be processed and the defocused image of the SAR image to be processed;

[0068] Step A3: calculating a second factor of the SAR image to be processed based on the SAR image to be processed and the defocused image of the SAR image to be processed;

[0069] Step A4: calculating an auxiliary factor of the SAR image to be processed based on the first factor and the second factor;

[0070] Step A5: Perform a two-dimensional inverse Fourier transform on the auxiliary factors to obtain a two-dimensional PSF estimation matrix of the SAR image to be processed.

[0071] In this embodiment, a defocused image of a SAR image to be processed is first obtained, and then based on the SAR image to be processed and the defocused image of the SAR image to be processed, the first factor and the second factor of the SAR image to be processed are respectively calculated, and the auxiliary factor of the SAR image to be processed is determined by the first factor and the second factor, and finally the auxiliary factor is inverse Fourier transformed to obtain a two-dimensional PSF estimation matrix of the SAR image to be processed. The two-dimensional PSF matrix solved by this method is closer to a regular two-dimensional Sinc function, so using the two-dimensional PSF matrix to participate in deconvolution calculation can obtain a better deconvolution result.

[0072] It should be noted that this embodiment introduces a modeling idea of ​​solving the image PSF based on the approximate greatest common divisor. First, a polynomial related to the SAR image to be processed is established, and then the auxiliary factor of the SAR image to be processed is obtained in the process of solving the greatest common divisor of the polynomial. The auxiliary factor is inverse Fourier transformed to obtain a two-dimensional PSF estimation matrix of the SAR image to be processed. The matrix is ​​closer to a regular two-dimensional Sinc function. Therefore, using the matrix to participate in deconvolution can obtain a more accurate deconvolution result.

[0073] Let's first clarify the definitions of the greatest common divisor and auxiliary factor:

[0074] Given two polynomials P(z) and Q(z), the two polynomials can be expressed as:

[0075] P(z)=p 0 +p 1 z+…+p n z n ;

[0076] Q(z)=q 0 +q 1 z+…+q n z n ;

[0077] Suppose the greatest common divisor of the polynomials P(z) and Q(z) is M(z), then there exist two polynomials C(z) and D(z) such that the following holds:

[0078]

[0079] Then C(z) and D(z) are the auxiliary factors of the polynomials P(z) and Q(z).

[0080] Given the definition of the greatest common divisor, the modeling and solution process of steps A1 to A5 are described in detail below:

[0081] In some embodiments, step A1 includes:

[0082] Read in the SAR image S1 to be processed and the related parameters of the SAR image to be processed; wherein the related parameters include the number of azimuth sampling points N of the SAR image to be processed a , the number of sampling points in the distance direction N r 、Phase error coefficient [Ep a , Ep r ], blur kernel size [h r ,h a ], penalty parameter μ, equilibrium constant λ and maximum number of iterations Iter max ;

[0083] Perform a two-dimensional Fourier transform on the SAR image S1 to be processed to obtain the first frequency domain image SF 1 , S.F. 1 =FFT2(S 1 ), where FFT2() represents a two-dimensional Fourier operation;

[0084] Add the phase error to the first frequency domain image SF1 to obtain the second frequency domain image SF containing the phase error 2 =SF 1 ⊙(E a ⊙E r );where E a is the azimuth error, E r is the distance error;

[0085]

[0086]

[0087] Among them, ⊙ represents the dot product of the matrix, that is, the corresponding elements are multiplied, and · represents the matrix multiplication;

[0088] For the second frequency domain image SF 2 Perform a two-dimensional inverse Fourier transform to obtain the defocused image S of the SAR image to be processed 2 , S 2 =IFFT2(SF 2 ), where IFFT2() represents a two-dimensional inverse Fourier operation.

[0089] In this embodiment, the SAR image S to be processed is 1 Perform frequency domain transformation, add phase error, and then obtain the defocused image S 2 , in order to facilitate the later establishment of the SAR image S to be processed 1 and the defocused image S 2 The polynomial is the basis for constructing polynomial equations in the process of solving auxiliary factors based on the greatest common divisor idea.

[0090] It should be noted that among the relevant parameters of the SAR image to be processed, the number of azimuth sampling points N a And the number of sampling points N in the distance direction r The phase error coefficient [Ep a , Ep r ], blur kernel size [h r ,h a ], penalty parameter μ, equilibrium constant λ and maximum number of iterations Iter max It can be set independently according to user requirements, but the selection of phase error needs to ensure that the SAR image is not over-focused, so as to ensure the similarity between the obtained defocused image and the SAR image to be processed. The blur kernel size needs to be determined according to the sidelobe size of the SAR image to be processed, and the penalty parameter and the equilibrium constant are adjusted according to the quality of the SAR image; the maximum number of iterations can be determined according to the empirical value, and then adjusted according to the image quality after the sidelobe suppression is obtained. It can be understood that the more iterations, the higher the image quality after sidelobe suppression, but it will take up more computer resources and increase the calculation time. Generally speaking, a maximum number of iterations of 10 times can obtain a better target image. This application does not make specific limitations on the relevant parameters of the SAR image to be processed.

[0091] In some embodiments, step A2 includes:

[0092] The SAR image to be processed and the defocused image of the SAR image to be processed are respectively subjected to range Fourier transform to obtain a first image and a second image; wherein each row vector in the first image can be expressed as Each row vector in the second image can be expressed as Where k = 1, 2, ..., N a , l = 1, 2, ..., N r ;

[0093] Calculate the auxiliary factor of each row vector in the first image separately The first factor A(k, l) of the SAR image to be processed is obtained; wherein,

[0094] The following takes the k-th row vector in the first image as an example to explain the calculation process of the auxiliary factor of the k-th row vector:

[0095] make Where l = 1, 2, ..., N r ;

[0096] Calculate g 1 (l) and g 2 (l) The Pei Shu matrix Bez(g 1 , g2 );

[0097] Let ∑·s=f; where, is the Pei Shu matrix Bez(g 1 , g 2 ) of (N a -h a )×(N a -h a ) of the main matrix, -f is the Pei Shu matrix Bez(g 1 , g 2 )No.(N a -h a +1)First in the column (N a -h a ) elements,

[0098] Using the singular value decomposition method to solve the formula

[0099] Calculate g 1 (l) cofactor Where J is an anti-diagonal matrix, the diagonal non-zero elements of an anti-diagonal matrix are 1, ⊙ represents the dot product of the matrix, that is, the corresponding elements are multiplied, and · represents matrix multiplication;

[0100] The above is the calculation process of the auxiliary factor of the k-th row vector. By using the above calculation method, the auxiliary factor of each row vector in the first image is calculated respectively, and the first factor A(k, l) of the SAR image to be processed can be obtained; wherein,

[0101] In some embodiments, step A3 includes:

[0102] Perform azimuth Fourier transform on the SAR image to be processed and the defocused image of the SAR image to be processed, respectively, to obtain a third image and a fourth image; wherein each row vector in the third image can be expressed as Each row vector in the fourth image can be expressed as Where k = 1, 2, ..., N a , l = 1, 2, ..., N r ;

[0103] Calculate the auxiliary factors of each row vector in the third image separately The second factor B(k, l) of the SAR image to be processed is obtained; wherein,

[0104] It should be noted that the calculation process of the second factor B(k, l) of the SAR image to be processed is the same as the calculation process of the first factor A(k, l), and only the processing order of rows and columns is exchanged.

[0105] In some embodiments, step A4 includes:

[0106] Combining the first factor A(k, l) and the second factor B(k, l), we get the first equation: A(k, l)a(k)-B(k, l)b(l)=0; where,

[0107] Integrate the first equation into the second equation: Γ·t=0; where,

[0108]

[0109] t=(a(0) … a(N a -1) b(0) … b(N r -1)) T ;

[0110] The singular value decomposition method is used to solve t, calculate a(k) and b(l), and obtain the auxiliary factors of the SAR image to be processed. Where k = 1, 2, ..., N a , l = 1, 2, ..., N r -1.

[0111] In some embodiments, step A5 includes:

[0112] Auxiliary factors for the SAR images to be processed Perform a two-dimensional inverse Fourier transform to obtain a two-dimensional PSF estimation matrix T of the SAR image to be processed; where, = 1, 2, ..., N a , l = 1, 2, ..., N r .

[0113] In this embodiment, the auxiliary factors of the SAR image to be processed are obtained by modeling and solving based on the greatest common divisor concept. Therefore, the two-dimensional PSF matrix solved by using the auxiliary factors is closer to the two-dimensional sinc function and closer to reality.

[0114] With respect to step 102, in some implementations, it includes:

[0115] The number of azimuth sampling points N based on the SAR image to be processed a And the number of sampling points N in the distance direction r , generate image Y; where the size of image Y is (N a +h a )×(N r+h r ), and the SAR image to be processed is located at the center of image Y. In the matrix corresponding to image Y, except for the pixels at the location of the SAR image to be processed, the other pixels are 0;

[0116] Construct the deconvolution function: y = Hx + n; where, is the one-dimensional vector generated by straightening the image Y, x is the deconvolution result to be calculated, n is the independent and identically distributed Gaussian white noise, H is the size of N a N r ×N a N r The block circulant matrix H is calculated by the two-dimensional PSF matrix T;

[0117] make Convert the deconvolution function into the following third-order function group:

[0118]

[0119]

[0120] Where N = N a ×N r , m=N a ×h r +N r ×h a +h a ×h r , is a one-dimensional vector generated by straightening the SAR image to be processed, z=zeros(m,1) is the zero vector, (a) j represents the jth element of vector a, D h and D v Represents the first-order derivative of the SAR image to be processed in azimuth and range directions, λ>0;

[0121] The third equation is iteratively calculated using the fast deconvolution method of the alternating direction multiplier method, that is, the following calculation processes are performed in sequence:

[0122] Step B1: Calculation

[0123] Step B2: Calculate z i+1 =2M z Hx i+1 -z i ; Among them, M z =[0,I d ], 0 is the zero vector, length is N, 1 d is a vector of all 1s, with a length of m;

[0124] Step B3: Calculation

[0125] Step B4: Calculate d i+1 =d i -(v i+1 -Dx i+1 );

[0126] Step B5: Calculation

[0127] Step B6: Update d i+1 for:

[0128] Update i=i+1, and repeat steps B1 to B6 until the maximum number of iterations Iter is reached. max ;

[0129] The deconvolution result x is output, and the convolution result x is restored to a two-dimensional image, which is the SAR image after sidelobe suppression.

[0130] In this embodiment, the deconvolution method is used to suppress the side lobes in the SAR image, which can effectively improve the image quality of the SAR image.

[0131] The sidelobe suppression effect of the SAR image sidelobe suppression method provided by the present application is described below with a specific embodiment:

[0132] Step 1: Use the parameters in Table 1 to simulate the SAR image;

[0133] Table 1 Figure 2 Relevant parameters of the SAR image shown

[0134]

[0135] The specific steps are as follows: Combined with the parameters in Table 1, 9 point targets are set in the simulation model. The position and amplitude of each point target are randomly generated, where the amplitude range is [0, 1]. Then, a SAR image is formed through SAR echo simulation and SAR imaging processing. Gaussian white noise is added to the SAR image to form a SAR image containing side lobes and Gaussian white noise. The positions of the 9 target points have been marked in the figure, as shown in Figure 2 shown.

[0136] It should be noted that, in this embodiment, the data may be used to generate a SAR image containing side lobes and Gaussian white noise, or a SAR image containing noise may be directly read in. In this case, there is no need to re-add noise.

[0137] Step 2: Set the phase error coefficient [Ep a , Ep r ], blur kernel size [h r ,ha ], penalty parameter μ, equilibrium constant λ, maximum number of iterations Iter max ;

[0138] Step 3: Combine the SAR image containing sidelobe noise given in step 1 and the phase error coefficient [Ep a , Ep r ], generating a defocused image;

[0139] Step 4: Combine the blur kernel size [h r ,h a ],calculate Figure 2 The first factor A of the SAR image shown;

[0140] Step 5: Combine the blur kernel size [h r ,h a ],calculate Figure 2 The second factor B of the SAR image shown;

[0141] Step 6: Combine the cofactors A and B calculated in steps 4 and 5 to calculate Figure 2 The two-dimensional PSF estimation matrix T of the SAR image shown is Figure 3 As shown;

[0142] Step 7: Combine the penalty parameter μ, equilibrium constant λ, and maximum number of iterations Iter given in step 2 max , and the two-dimensional PSF estimation matrix T of step six, and iteratively calculate the image deconvolution result x using a fast deconvolution method based on an alternating direction multiplier method;

[0143] The convolution result x is restored to a two-dimensional image and the Figure 2 SAR image after sidelobe suppression, and the positions of the 9 target points are marked in the figure, as shown in the figure. Figure 4 shown.

[0144] It should be noted that, in order to facilitate the location of the target point, this embodiment performs a noise SAR image ( Figure 2 ) and SAR image after sidelobe suppression ( Figure 4 ) are interpolated, and the interpolation multiple is 50 times. In addition, for ease of viewing, this embodiment Figure 2 The two-dimensional PSF estimation matrix T of the noisy SAR image shown in Figure 1 is interpolated and converted into a dB graph, as shown in Figure 2. Figure 3 As shown in the figure, it can be seen that the two-dimensional PSF estimation matrix T using the method proposed by the present invention is approximately a two-dimensional sinc function, which is consistent with the actual situation. Figure 4 It can be clearly seen from the sidelobe suppression result diagram shown that the method proposed in the present invention can effectively filter out noise and restore the real information of the scene.

[0145] Figure 5 and Figure 6 They are Figure 2 The sidelobe suppression result profile of target point 5 in the noisy SAR image in the range and azimuth directions is shown. It can be seen from the figure that the method proposed in the present invention can effectively suppress the sidelobe information while maintaining the main lobe resolution. Compared with the windowing method and the SVA method, the sidelobe suppression performance is the best.

[0146] In summary, the SAR image sidelobe suppression method provided by the present invention can effectively filter out noise and has good sidelobe suppression performance.

[0147] like Figure 7 , Figure 8 As shown, an embodiment of the present invention provides a SAR image sidelobe suppression device. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. From the hardware level, Figure 7 As shown, it is a hardware architecture diagram of an electronic device in which a SAR image sidelobe suppression device provided by an embodiment of the present invention is located. Figure 7 In addition to the processor, memory, network interface, and non-volatile memory shown, the electronic device in the embodiment may also include other hardware, such as a forwarding chip responsible for processing messages, etc. Taking software implementation as an example, Figure 8 As shown, as a device in a logical sense, the CPU of the electronic device in which it is located reads the corresponding computer program in the non-volatile memory into the internal memory and runs it.

[0148] like Figure 8 As shown, a SAR image sidelobe suppression device provided by this embodiment includes:

[0149] A calculation module 300 is used to calculate a two-dimensional PSF estimation matrix of a SAR image to be processed;

[0150] The solution module 302 is used to solve the deconvolution function based on the two-dimensional PSF estimation matrix and the relevant parameters of the SAR image to be processed, and output the deconvolution result; wherein the deconvolution result can be restored to the SAR image after sidelobe suppression, and the relevant parameters include the number of sampling points in the azimuth direction, the number of sampling points in the range direction, the phase error coefficient, the blur kernel size, the penalty parameter, the equilibrium constant and the maximum number of iterations of the SAR image to be processed, the azimuth direction is the direction of the radar flight track, and the range direction is the direction along the radar wave emission.

[0151] In the embodiment of the present invention, the calculation module 300 may be used to execute step 100 in the above method embodiment, and the solution module 302 may be used to execute step 102 in the above method embodiment.

[0152] In one embodiment of the present invention, the calculation module 300 is used to perform the following operations:

[0153] Acquire a defocused image of the SAR image to be processed;

[0154] Calculating a first factor of the SAR image to be processed based on the SAR image to be processed and the defocused image of the SAR image to be processed;

[0155] Calculating a second factor of the SAR image to be processed based on the SAR image to be processed and the defocused image of the SAR image to be processed;

[0156] Based on the first factor and the second factor, an auxiliary factor of the SAR image to be processed is calculated;

[0157] Perform a two-dimensional inverse Fourier transform on the auxiliary factors of the SAR image to be processed to obtain a two-dimensional PSF estimation matrix of the SAR image to be processed.

[0158] In one embodiment of the present invention, the solution module 302 is used to perform the following operations:

[0159] The number of azimuth sampling points N based on the SAR image to be processed a And the number of sampling points N in the distance direction r , generate image Y; where the size of image Y is (N a +h a )×(N r +h r ), and the SAR image to be processed is located at the center of image Y. In the matrix corresponding to image Y, except for the pixels at the location of the SAR image to be processed, all other pixels are 0;

[0160] Construct the deconvolution function: y = Hx + n; where, is the one-dimensional vector generated by straightening the image Y, x is the deconvolution result to be calculated, n is the independent and identically distributed Gaussian white noise, H is the size of N a N r ×N a N r The block circulant matrix H is calculated by the two-dimensional PSF matrix T;

[0161] make Convert the deconvolution function into the following third-order function group:

[0162]

[0163]

[0164] Where N = N a ×N r , m=Na ×h r +N r ×h a +h a ×h r , is a one-dimensional vector generated by straightening the SAR image to be processed, z=zeros(m,1) is the zero vector, (a) j represents the jth element of vector a, D h and D v Represents the first-order derivative of the SAR image to be processed in azimuth and range directions, λ>0;

[0165] The third equation is iteratively calculated using the fast deconvolution method of the alternating direction multiplier method until the maximum number of iterations Iter is reached. max , output the deconvolution result x.

[0166] It is to be understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on the sidelobe suppression device for SAR images. In other embodiments of the present invention, a sidelobe suppression device for SAR images may include more or fewer components than those illustrated, or combine certain components, or separate certain components, or arrange the components differently. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0167] The information interaction, execution process and other contents between the modules in the above-mentioned device are based on the same concept as the embodiment of the method of the present invention. For the specific contents, please refer to the description in the embodiment of the method of the present invention, and no further description is given here.

[0168] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, a sidelobe suppression method for a SAR image in any embodiment of the present invention is implemented.

[0169] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the processor executes a sidelobe suppression method for a SAR image in any embodiment of the present invention.

[0170] Specifically, a system or device equipped with a storage medium can be provided, on which software program code that implements the functions of any of the above-mentioned embodiments is stored, and a computer (or CPU or MPU) of the system or device can be enabled to read and execute the program code stored in the storage medium.

[0171] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0172] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.

[0173] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling the system operating on the computer to complete part or all of the actual operations based on the instructions of the program code.

[0174] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion module connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or expansion module is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0175] It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements limited by the statement "comprise a ..." do not exclude the existence of other identical factors in the process, method, article or equipment including the elements.

[0176] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.

[0177] 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 sidelobe suppression method for SAR images, characterized in that: include: Calculate the two-dimensional PSF estimation matrix of the SAR image to be processed; Based on the two-dimensional PSF estimation matrix and the relevant parameters of the SAR image to be processed, a deconvolution function is solved and a deconvolution result is output; wherein the deconvolution result can be restored to a SAR image after sidelobe suppression, and the relevant parameters include the number of sampling points in the azimuth direction, the number of sampling points in the range direction, the phase error coefficient, the blur kernel size, the penalty parameter, the equilibrium constant and the maximum number of iterations of the SAR image to be processed, the azimuth direction is the direction of the radar flight track, and the range direction is the direction along which the radar wave is emitted; The step of calculating a two-dimensional PSF estimation matrix of a SAR image to be processed includes: Acquiring a defocused image of the SAR image to be processed; Performing range Fourier transform on the SAR image to be processed and the defocused image of the SAR image to be processed respectively to obtain a first image and a second image; calculating the auxiliary factor of each row vector in the first image respectively to obtain a first factor of the SAR image to be processed; Performing azimuth Fourier transform on the SAR image to be processed and the defocused image of the SAR image to be processed respectively to obtain a third image and a fourth image; calculating the auxiliary factor of each row vector in the third image respectively to obtain a second factor of the SAR image to be processed; Based on the first factor and the second factor, calculating an auxiliary factor of the SAR image to be processed; Performing a two-dimensional inverse Fourier transform on the auxiliary factors of the SAR image to be processed to obtain a two-dimensional PSF estimation matrix of the SAR image to be processed; The step of solving a deconvolution function based on the two-dimensional PSF matrix and relevant parameters of the SAR image to be processed and outputting a deconvolution result comprises: Based on the number of azimuth sampling points N of the SAR image to be processed a And the number of sampling points N in the distance direction r , generate image Y; wherein the size of the image Y is (N a +h a )×(N r +h r ), and the SAR image to be processed is located at the center of the image Y, and in the matrix corresponding to the image Y, except for the pixel at the location of the SAR image to be processed, the other pixels are 0; Construct the deconvolution function: y = Hx + n; where, is a one-dimensional vector generated by straightening the image Y, x is the deconvolution result to be calculated, n is independent and identically distributed Gaussian white noise, H is a vector with a size of N a N r ×N a N r A block circulant matrix H is calculated by the two-dimensional PSF matrix T; make The deconvolution function is transformed into the following third set of equations: Where N = N a ×N r , m=N a ×h r +N r ×h a +h a ×h r , is a one-dimensional vector generated by straightening the SAR image to be processed, z=zeros(m, 1) is a zero vector, (D h x) j Represents vector D h The jth element of x, (D v x) j Represents vector D v The jth element of x, D h and D v Represents the first-order derivatives of the SAR image to be processed in azimuth and range directions, λ>0; The third equation is iteratively calculated using the fast deconvolution method of the alternating direction multiplier method until the maximum number of iterations Iter is reached. max , output the deconvolution result x.

2. The method according to claim 1, characterized in that: The step of acquiring the defocused image of the SAR image to be processed comprises: Read in the SAR image to be processed and related parameters of the SAR image to be processed; wherein the related parameters include the number of azimuth sampling points N of the SAR image to be processed a , the number of sampling points in the distance direction N r 、Phase error coefficient [Ep a , Ep r ], blur kernel size [h r ,h a ], penalty parameter μ, equilibrium constant λ and maximum number of iterations Iter max ; Performing a two-dimensional Fourier transform on the SAR image to be processed to obtain a first frequency domain image; Adding a phase error to the first frequency domain image to obtain a second frequency domain image containing the phase error; Perform a two-dimensional inverse Fourier transform on the second frequency domain image to obtain a defocused image of the SAR image to be processed.

3. The method according to claim 2, characterized in that Each row vector in the first image can be expressed as Each row vector in the second image can be expressed as Where k = 1, 2, ..., N a , l = 1, 2, ..., N r ; The auxiliary factor of each row vector in the first image is The first factor is A(k, l); where 4. The method according to claim 3, characterized in that Each row vector in the third image can be expressed as Each row vector in the fourth image can be expressed as Where k = 1, 2, ..., N a , l = 1, 2, ..., N r ; The auxiliary factor of each row vector in the third image is The second factor is B(k, l); where 5. The method according to claim 4, characterized in that The step of calculating the auxiliary factor of the SAR image to be processed based on the first factor and the second factor includes: The first factor A(k, l) and the second factor B(k, l) are combined to obtain a first equation: A(k, l)a(k)-B(k, l)b(l)=0; wherein, The first equation is integrated into the second equation: Γ·t=0; where, t=(a(0) … a(N a -1) b(0) … b(N r -1)) T ; The singular value decomposition method is used to solve the t, calculate the a(k) and the b(l), and obtain the auxiliary factor of the SAR image to be processed. Where k = 1, 2, ..., N a , l = 1, 2, ..., N r -1.

6. A SAR image sidelobe suppression device, characterized in that: The sidelobe suppression method for the SAR image according to claim 1, wherein the device comprises: A calculation module, used for calculating a two-dimensional PSF estimation matrix of a SAR image to be processed; A solution module is used to solve the deconvolution function based on the two-dimensional PSF estimation matrix and the relevant parameters of the SAR image to be processed, and output the deconvolution result; wherein the deconvolution result can be restored to the SAR image after sidelobe suppression, and the relevant parameters include the number of sampling points in the azimuth direction, the number of sampling points in the range direction, the phase error coefficient, the blur kernel size, the penalty parameter, the equilibrium constant and the maximum number of iterations of the SAR image to be processed, the azimuth direction is the direction of the radar flight track, and the range direction is the direction along the radar wave emission.

7. An electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the method according to any one of claims 1 to 5.

Images