Rapid two-dimensional clar detection method for sar imaging recognition

Abstract

The invention discloses a rapid two-dimensional cfu detection method for sar imaging recognition, relates to the field of imaging, and aims to solve the problems that a traditional cfu algorithm is high in time complexity and insufficient in detection precision, and needs a large calculation amount and long operation time. The method comprises the following steps: dividing a sar two-dimensional amplitude image into four areas; calculating the area of a detection window annular region of the central region, and calculating a clar threshold value based on the area of the detection window annular region; performing sliding window processing on the two-dimensional integrogram of the central area to obtain a detection window sum value of each detection point; calculating the cfu value of each detection point based on the cfu threshold value and the detection window sum value; comparing each detection point of the sar two-dimensional amplitude image of the central area with the cfu value of each detection point to obtain a binary image of the central area; and performing clar detection on the four-corner boundary region, the left-right boundary region and the upper-lower boundary region to obtain binary images of the four-corner boundary region, the left-right boundary region and the upper-lower boundary region. According to the method, the detection accuracy is improved, and the clar calculation time is shortened.

Description

Technical Field

[0001] This document relates to the field of imaging, and in particular to a fast two-dimensional CFAR detection method for SAR imaging recognition. Background Technology

[0002] SAR is an active imaging sensor-based synthetic aperture radar. This radar features all-weather, all-time operation and strong penetration capabilities, making it widely used in both civilian and military fields. In multi-target scenarios, the radar system can simultaneously measure the target's range and radial velocity. After receiving the echo signals and performing correlation processing, a two-dimensional range-Doppler spectrum matrix is ​​obtained. During the detection phase, a constant false alarm rate (CFAR) detector is typically used to identify regions of interest (ROIs) of potential targets within a large image area. The purpose of CFAR design is to provide a detection threshold that relatively avoids the influence of background noise, clutter, and interference variations. Commonly used detection methods include CA-CFAR, OS-CFAR, and dual-parameter CFAR, with dual-parameter CFAR being the most widely used. However, for high-resolution SAR images, the size of the rectangular window is often quite large. Dual-parameter CFAR uses this sliding window to detect each pixel and calculates the mean and variance of the background, which requires significant computation and a long processing time.

[0003] Therefore, a fast two-dimensional CFAR detection method based on SAR imaging is proposed. Summary of the Invention

[0004] This specification provides a fast two-dimensional CFAR detection method for SAR imaging recognition, which solves the problems of high time complexity and insufficient detection accuracy of traditional CFAR algorithms. For example, two-dimensional CFAR detection algorithms use a sliding window to detect each pixel and calculate the mean and variance of the background, thus requiring a large amount of computation and a long processing time. The method includes:

[0005] The two-dimensional distance Doppler SAR image is converted into a two-dimensional SAR amplitude image. The two-dimensional SAR amplitude image is then divided into four regions: the central region, the four corner boundary regions, the left and right boundary regions, and the upper and lower boundary regions.

[0006] The Doppler value range of the central region is obtained based on the distance value range of the central region; the area of ​​the annular region of the detection window in the central region is calculated, and the CFAR threshold is calculated based on the area of ​​the annular region of the detection window;

[0007] The SAR two-dimensional amplitude image of the central region is processed into an integral image to obtain a two-dimensional integral image; the two-dimensional integral image is processed into a sliding window to obtain the detection window and value of each detection point; the CFAR value of each detection point is calculated based on the CFAR threshold and the detection window and value; the binary image of the central region is obtained by comparing each detection point of the SAR two-dimensional amplitude image with the CFAR value of each detection point.

[0008] CFA detection is performed on the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions to obtain binary maps of the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions, respectively.

[0009] In a preferred embodiment, obtaining the Doppler value range of the central region based on the distance value range of the central region includes:

[0010] The range of distance values ​​for the central region is:

[0011] [(Pro_Dis_L+Dest_Dis_L-1),M-(Pro_Dis_L+Dest_Dis_L-1)];

[0012] Where Pro_Dis_L is the half-window length of the protection window in the distance dimension of the central region, and M is the length of the distance dimension.

[0013] Dest_Dis_L is the half-window length of the detection window at the center region;

[0014] The range of Doppler values ​​in the central region is:

[0015] [(Pro_Dpl_L+Dest_Dpl_L-1),N-(Pro_Dpl_L+Dest_Dpl_L-1)];

[0016] Where Pro_Dpl_L is the half-window length of the Doppler protection window, N is the Doppler length, and Dest_Dpl_L is the half-window length of the Doppler detection window.

[0017] In a preferred embodiment, the method for calculating the CFAR threshold includes:

[0018] First, calculate the area of ​​the annular region of the detection window:

[0019] cen_win_S=((Det_L+Pro_L)×2+1)×((Det_L+Pro_L)×2+1)-(Pro_L×2+1)×(Pro_L×2+1);

[0020] Where cen_win_S is the area of ​​the annular region of the detection window, Det_L is the length of the Doppler detection window, and Pro_L is the length of the Doppler protection window;

[0021] Then, the cfar threshold is calculated based on the area of ​​the annular region of the detection window;

[0022]

[0023] Wherein, CFAR_PFA is the CFAR constant false alarm factor, cen_win_S is the area of ​​the annular region of the detection window, and cen_pfa is the CFAR threshold.

[0024] In a preferred embodiment, performing a sliding window process on the two-dimensional integral image to obtain the detection window and value for each detection point includes:

[0025] The two-dimensional integral image is processed by sliding windowing to obtain the outer window sum for each detection point:

[0026] sum1 i =rb i +lu i -ru i -lb i ;

[0027] Among them, rb i Let lu be the value of the lower right point of the outer window of the i-th detection point in the two-dimensional integral image. i Let ru be the value of the top-left point of the outer window of the i-th detection point in the two-dimensional integral graph. i Let lb be the value of the upper right point of the outer window of the i-th detection point in the two-dimensional integral graph. i The sum1 is the value of the lower left point of the outer window of the i-th detection point in the two-dimensional integral graph; i Let the sum of the outer windows of the i-th detection point be ;

[0028] The two-dimensional integral image is subjected to sliding window processing to obtain the inner window sum of each detection point;

[0029] The detection window and value for each detection point are:

[0030] sum i =sum1 i -sum2 i ;

[0031] Where, sum i Let sum1 be the sum of the detection window values ​​at the i-th detection point, sum2 be the sum of the outer window values ​​at the i-th detection point, and sum3 be the sum of the outer window values ​​at the i-th detection point. i Let be the sum of the inner windows of the i-th detection point.

[0032] In a preferred embodiment, calculating the CFAR value for each detection point based on the CFAR threshold and the detection window sum value includes:

[0033] cfar_y i =sum i ×cen_pfa

[0034] Where, sum i Here, cen_pfa represents the detection window and value at the i-th detection point in the central region, and cen_pfa represents the CFAR threshold at the i-th detection point in the central region; cfar_y i is the cfar value of the i-th detection point in the central region.

[0035] In a preferred embodiment, performing CFA detection on the four corner boundary regions includes:

[0036] The size of the four corner boundary regions is cfar_L:

[0037] cfar_L=((Det_L+Pro_L)*2+1)×((Det_L+Pro_L)*2+1);

[0038] Where Det_L is the length of the detection window and Pro_L is the length of the protection window;

[0039] Calculate the CFA detection values ​​of the four corner points of the four-corner boundary region based on the CFA value of each detection point in the central region;

[0040] lu_th=sum [0] *cen_pfa;

[0041] ru_th=sum [N-2×(Det_l+Pro_l)-1] ×cen_pfa;

[0042] lb_th = sum [(M-2×(Det_l+Pro_l-1))*N] ×cen_pfa;

[0043] rb_th = sum [(M-2×(Det_l+Pro_l) -1)*N+N-2×(Det_l+Pro_l)-1]×cen_pfa;

[0044] Where, lu_th is the CFAR detection value of the top-left corner, ru_th is the CFAR detection value of the top-right corner, lb_th is the CFAR detection value of the bottom-left corner, and rb_th is the CFAR detection value of the bottom-right corner; M represents the length of the range direction of the two-dimensional range-Doppler SAR image, and N represents the length of the azimuth direction of the two-dimensional range-Doppler SAR image; Det_L is the length of the detection window, Pro_L is the length of the protection window; sum [0] For the outer window of the 0th detection point, sum [N-2×(Det_l+Pro_l)-1] The sum of the outer window values ​​at the (N-2×(det_l+pro_l)-1)th detection point is given by sum.[(M-2×(Det_l+Pro_l-1))*N] Let sum be the sum of the outer windows of the (M-2×(Det_l+Pro_l)-1)*Nth detection point. [(M-2×(Det_l+Pro_l)-1)*N+N-2×(Det_l+Pro_l)-1] The sum of the outer window of the (M-2×(Det_l+Pro_l)-1)*N+N-2×(Det_l+Pro_l)-1 detection point.

[0045] In a preferred embodiment, CFAR detection of the four corner boundary regions further includes:

[0046] Each detection point in the upper left corner is compared with the cfar detection value lu_th of the upper left corner point. If the detection point is greater than lu_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the upper left corner.

[0047] Each detection point in the upper right corner is compared with the cfar detection value ru_th of the upper right corner point. If the detection point is greater than ru_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the upper right corner.

[0048] Each detection point in the lower left corner is compared with the cfar detection value lb_th of the lower left corner point. If the detection point is greater than lb_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the lower left corner.

[0049] Each detection point in the lower right corner is compared with the cfar detection value rb_th of the lower right corner point. If the detection point is greater than rb_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the lower right corner.

[0050] In a preferred embodiment, performing CFA detection on the left and right boundary regions includes:

[0051] The left boundary region detection window uses the right half-window CFAR detection, while the right boundary region detection window uses the left half-window CFAR detection.

[0052] The sliding window size used for CFA detection of the left and right boundary regions is:

[0053] dpl_win_S=(Det_L+Pro_L+1)×((Det_L+Pro_L)×2+1);

[0054] Where dpl_win_S is the size of the sliding window used for CFA detection of the left and right boundary regions, Det_L is the length of the detection window, and Pro_L is the length of the protection window.

[0055] Compare all detection points within the left and right boundary regions with the corresponding CFA detection values. If a value is greater than the detection value, set the value to 1; otherwise, set it to 0. This will yield the binary map of the upper region.

[0056] In a preferred embodiment, performing CFA detection on the left boundary region specifically includes:

[0057] The sum of the outer windows of the b-th detection point in the first row of the left boundary region is:

[0058] huge_win_sum[b]=huge_win_sum[b-1]-huge_sum_col[b]+huge_sum_col[b+Pro_L+Det_L];

[0059] Where huge_win_sum[b] is the sum of the outer windows of the b-th sliding window block in the first row, huge_win_sum[b-1] is the sum of the outer windows of the (b-1)-th sliding window block in the first row, huge_sum_col[b] is the sum of the outer windows of the detection points in the b-th column of the first row, and huge_sum_col[b+Pro_L+Det_L] is the sum of the outer windows of the detection points in the b+pro_L+Det_L column of the first row;

[0060] The sum of the inner windows at the b-th detection point in the first row of the left boundary region is: -

[0061] pro_win_sum[b]=pro_win_sum[b-1]+small_sum_col[b+Pro_L]-sumall_sum_col[b];

[0062] Where pro_win_sum[b] is the inner window sum of the b-th detection point in the first row, pro_win_sum[b-1] is the inner window sum of the (b-1)-th detection point in the first row, small_sum_col[b+Pro_L] is the inner window sum of the b+Pro_L-th column in the first row, and sumall_sum_col[b] is the inner window sum of the b-th column in the first row;

[0063] The detection window sum value corresponding to the b-th detection point in the first row within the left boundary region is:

[0064] det_win_sum[b]=huge_win_sum[b]-pro_win_sum[b];

[0065] Where huge_win_sum[b] is the outer window sum of the b-th detection point in the first row; pro_win_sum[b] is the inner window sum of the b-th detection point in the first row; and det_win_sum[b] is the detection window sum value corresponding to the b-th detection point in the first row;

[0066] The sum of the outer windows of the detection points in the q-th row and k-th column of the left boundary region, excluding the first row, is:

[0067] huge_sum_col[k]=huge_sum_col[k-1]-cfar_sb[(q-Pro_L-Dis_L-1)×N+k]+cfar_sb[(q+Pro_L+Dis_L)×N+k];

[0068] huge_sum_col[k] is the sum of the outer windows in the q-th row and k-th column excluding the first row, huge_sum_col[k-1] is the sum of the outer windows in the q-th row and (k-1)-th column excluding the first row, cfar_sb[(q-Pro_L-Dis_L-1)×N+k] is the pixel value of the (q-Pro_L-Dis_L-1)×N+k-th point on the SAR image of the left boundary region, where N represents the length of the azimuth direction of the two-dimensional distance Doppler SAR image; cfar_sb(q+Pro_L+Dis_L)×N+k is the pixel value of the (q-Pro_L-Dis_L-1)×N+k-th point on the SAR image of the left boundary region.

[0069] The sum of the inner windows of the detection points in the q-th row and k-th column of the left boundary region, excluding the first row, is:

[0070] small_sum_col[k]=small_sum_col[k-1]-cfar_sb[(q-Pro_L-1)×N+k]+cfar_sb[(q+Pro_L)×N+k];

[0071] Where, small_sum_col[k-1] is the sum of the inner windows in the q-th row and k-th column excluding the first row, small_sum_col[k] is the sum of the inner windows in the q-th row and k-th column excluding the first row, cfar_sb is the length of the azimuth direction of the SAR image N, representing the two-dimensional distance Doppler SAR image; cfar_sb[(q-Pro_L-1)×N+k] is the (q-Pro_L-1)×N+k pixel on the SAR image of the left boundary region; cfar_sb[(q+Pro_L)×N+k] is the [(q+Pro_L)×N+k] pixel on the SAR image of the left boundary region;

[0072] The detection window sum value corresponding to the detection point in the q-th row and k-th column of the left boundary region (excluding the first row) is:

[0073] det_win_sum[k]=huge_sum_col[k]-small_sum_col[k];

[0074] The half-window detection area of ​​the left boundary region is:

[0075] s_det_dpl=((det_L+Pro_L)×2+1)×(det_L+Pro_L+1)-(Pro_L+1)×Pro_L×2+1);

[0076] The CFA detection values ​​for the detection points in the left half of the region are:

[0077]

[0078] Among them, dpl_hf_pfa b is the CFAR detection value of the b-th detection point within the left boundary region; cfar_sb is the detection point of the SAR image within the left boundary region; S_det_dpl is the area of ​​the left boundary region.

[0079] Compare all detection points within the left boundary region with the corresponding CFAR detection value. If a detection point is larger than the CFAR detection value, set the value to 1; otherwise, set it to 0. This will give you the binary map of the upper region.

[0080] In a preferred embodiment, performing CFA detection on the upper and lower boundary regions includes:

[0081] The upper boundary region detection window uses the lower half-window CFAR detection, while the lower boundary region detection window uses the upper half-window CFAR detection.

[0082] The sizes of the sliding windows in the upper and lower boundary areas are:

[0083] cfar_dis_win=(Det_L+Pro_L+1)×((Det_L+Pro_L)×2+1);

[0084] Where Det_L is the detection window length and Pro_L is the protection window length;

[0085] Adding the sums of each column within the first sliding window of the first row will give you the sum of the outer window of the first sliding window in the first row: sum_win_dis1;

[0086] The sum of the outer windows of the first sliding window in the first row is used to obtain the sum of the outer windows of the second sliding window; the sum of the outer windows of the m-th sliding window in the first row is used to obtain the sum of the outer windows of the (m+1)-th sliding window in the first row.

[0087] The sum of the outer windows of the (m+1)th sliding window in the first row is:

[0088] sum_win_dis m+1 =sum_win_dis m +dis_sum_col[m+Det_L+Pro_L+1]-dis_sum_col[m-Det_L+Pro_L];

[0089] Among them, sum_win_dis m+1 The sum of the outer windows of the (m+1)th sliding window in the first row of the upper boundary region, sum_win_dis m The sum of the outer windows of the m-th sliding window in the first row of the upper boundary region is given by dis_sum_col[m+Det_L+Pro_L+1], where m+Det_L+Pro_L+1 is the sum of the outer windows of the m-th sliding window in the first row of the upper boundary region.

[0090] The sum of the outer window in column 1, dis_sum_col[m-Det_L+Pro_L] is the sum of the outer window in column m-Det_L+Pro_L within the first row of the sliding window block in the upper boundary region;

[0091] The sum of the protection windows of the first sliding window in the first row is used to obtain the sum of the protection windows of the second sliding window in the first row; the sum of the protection windows of the second sliding window in the first row is used to obtain the sum of the protection windows of the third sliding window in the first row; the sum of the protection windows of the m-th sliding window in the first row is used to obtain the sum of the (m+1)-th protection windows in the first row; the sum of the protection windows corresponding to the m-th detection point in the first row is: sum_small_win_dis m ;

[0092] The sum of the outer windows of the detection points in the r-th row and t-th column, excluding the first row of the upper boundary region, is:

[0093] dis_sum_col[t]=dis_sum_col[t-1]-cfar_sm[(r-1)×N+t]+cfar_sm[(r+Det_L+Pro_L)*N+t];

[0094] Where, dis_sum_col[t] is the sum of the outer windows in the r-th row and t-th column excluding the first row of the upper boundary region, dis_sum_col[t-1] is the sum of the outer windows in the r-th row and t-1-th column excluding the first row of the upper boundary region, cfar_sm[(r-1)×N+t] is the pixel value of the (r-1)×N+t-th point on the SAR image of the upper boundary region, and cfar_sm[(r+Det_L+Pro_L)*N+t] is the pixel value of the (r+Det_L+Pro_L)×N+t-th point on the SAR image of the upper boundary region;

[0095] The sum of the inner windows of the detection points in the r-th row and t-th column, excluding the first row of the upper boundary region, is:

[0096] dis_sum_small_col[t]=dis_sum_small_col[t-1]-cfar_sm[(r-1)×N+t]+cfar_sm[(r+Det_L+Pro_L)*N+t];

[0097] Where, dis_sum_small_col[t] is the inner window sum of the r-th row and t-th column excluding the first row, dis_sum_small_col[t-1] is the inner window sum of the r-th row and t-1-th column excluding the first row, cfar_sm[(r-1)×N+t] is the pixel value of the (r-1)×N+t-th point on the SAR image of the upper boundary region, and cfar_sm[(r+Det_L+Pro_L)*N+t] is the pixel value of the (r+Det_L+Pro_L)×N+t-th point on the SAR image of the upper boundary region;

[0098] The detection window and value sum_det_dis of the detection points in the first row of the upper boundary region m for:

[0099] sum_det_dis m =sum_win_dis m -sum_samll_win_dis m ;

[0100] The detection window and value sum_det_dis of the detection points excluding the first row in the upper boundary region t for:

[0101] sum_det_dis t =dis_sum_small_col[t]-sum_det_dis t ;

[0102] The half-window detection area of ​​the upper boundary region is:

[0103] S_det_dis=((det_L+Pro_L)×2+1)×(det_L+Pro_L+1-(Pro_L+1)×(Pro_L×2+1);

[0104] The CFA detection value of the detection points in the upper boundary region is:

[0105]

[0106] Where cfar_sm is the pixel value of the m-th point on the SAR image of the upper boundary region; dis_hf_pfa m S_det_dis is the CFA detection value of the m-th detection point in the upper boundary region; S_det_dis is the half-window detection area of ​​the upper boundary region.

[0107] For any detection point in the upper or lower boundary region, if the amplitude value of the detection point is greater than its corresponding CFA detection value, then the detection point is set to 1; otherwise, it is set to 0.

[0108] The above-described at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects:

[0109] This invention addresses the issue of 2D CFAR detection in SAR images. To overcome the shortcomings of traditional CFAR detection methods, such as insufficient accuracy and high time complexity, this invention proposes a fast 2D CFAR detection algorithm for SAR imaging. The amplitude map of the 2D SAR image is divided into four regions. Different algorithms are used to calculate the CFAR threshold for each region. The pixel values ​​of detection points in different regions are compared with the detection threshold for that region, ultimately obtaining a binary image. In implementation, an integral image and recursive approach are used to detect points in each region. This design improves detection accuracy and reduces CFAR computation time. Attached Figure Description

[0110] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0111] Figure 1 This is a schematic diagram of a fast two-dimensional CFAR detection method for SAR imaging recognition provided in one embodiment of this specification;

[0112] Figure 2 This is a schematic diagram of a four-region division method for SAR imaging provided in an embodiment of the present invention;

[0113] Figure 3 This is a schematic diagram illustrating the basic definition of a two-dimensional CFAR sliding window according to an embodiment of the present invention;

[0114] Figure 4 This is a schematic diagram illustrating the definition of a two-dimensional CFAR detection point according to an embodiment of the present invention;

[0115] Figure 5 This is a schematic diagram of the detection windows for various regions in a two-dimensional CFAR according to an embodiment of the present invention. Detailed Implementation

[0116] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0117] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0118] Figure 1 A fast two-dimensional CFAR detection method for SAR imaging identification is provided in one embodiment of this specification. The method may specifically include the following steps:

[0119] The two-dimensional distance Doppler SAR image is converted into a two-dimensional SAR amplitude image. The two-dimensional SAR amplitude image is then divided into four regions: the central region, the four corner boundary regions, the left and right boundary regions, and the upper and lower boundary regions.

[0120] The Doppler value range of the central region is obtained based on the distance value range of the central region; the area of ​​the annular region of the detection window in the central region is calculated based on the Doppler value range of the central region; and the CFAR threshold is calculated based on the area of ​​the annular region of the detection window.

[0121] The SAR two-dimensional amplitude image of the central region is processed into an integral image to obtain a two-dimensional integral image; the two-dimensional integral image is processed into a sliding window to obtain the detection window and value of each detection point; the CFAR value of each detection point is calculated based on the CFAR threshold and the detection window and value; the binary image of the central region is obtained by comparing each detection point of the SAR two-dimensional amplitude image with the CFAR value of each detection point.

[0122] CFA detection is performed on the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions to obtain binary maps of the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions, respectively.

[0123] The following is a detailed description of the fast two-dimensional CFAR detection method for SAR imaging recognition of the present invention; the SAR image CFAR_si[M×N] is a two-dimensional amplitude image with a size of M*N, where the distance upward length is M and the Doppler upward length is N.

[0124] The length of the protection unit of the SAR image is Pro_L, the length of the detection unit of the SAR image is Det_L, and the constant false alarm factor is CFAR_PFA. The constant false alarm factor is used to select the threshold of CFAR.

[0125] The two-dimensional distance Doppler SAR image is converted into a two-dimensional SAR amplitude image. The two-dimensional SAR amplitude image is then divided into four regions: the central region, the four corner boundary regions, the left and right boundary regions, and the upper and lower boundary regions.

[0126] like Figure 2 As shown, the red area is the central region. The Doppler value range of the central region is obtained based on the distance value range of the central region. The area of ​​the detection window annular region of the central region is calculated based on the Doppler value range of the central region, and the CFAR threshold is calculated based on the area of ​​the detection window annular region.

[0127] In this embodiment, obtaining the Doppler range of the central region based on the distance range of the central region includes:

[0128] The range of values ​​for the distance to the central region is:

[0129] [(Pro_Dis_L+Dest_Dis_L-1),M-(Pro_Dis_L+Dest_Dis_L-1)];

[0130] Where Pro_Dis_L is the half-window length of the protection window in the distance dimension of the central region, and M is the length of the distance dimension.

[0131] Dest_Dis_L is the half-window length of the detection window at the center region;

[0132] The Doppler range of the central region is calculated based on the range of distance values ​​in the central region.

[0133] [(Pro_Dpl_L+Dest_Dpl_L-1),N-(Pro_Dpl_L+Dest_Dpl_L-1)];

[0134] Where Pro_Dpl_L is the half-window length of the Doppler protection window, N is the Doppler length, and Dest_Dpl_L is the half-window length of the Doppler detection window.

[0135] In this embodiment, the method for calculating the CFAR threshold includes:

[0136] First, calculate the area of ​​the annular region of the detection window based on the Doppler value range of the central region:

[0137] cen_win_S=((Det_L+Pro_L)×2+1)×((Det_L+Pro_L)×2+1)-(Pro_L×2+1)×(Pro_L×2+1);

[0138] Where cen_win_S is the area of ​​the annular region of the detection window, Det_L is the width of the annular detection window, and Pro_L is the width of the protection window;

[0139] Then, the cfar threshold is calculated based on the area of ​​the annular region of the detection window;

[0140]

[0141] Wherein, CFAR_PFA is the CFAR constant false alarm factor, cen_win_S is the area of ​​the annular region of the detection window, and cen_pfa is the CFAR threshold.

[0142] The SAR two-dimensional amplitude image of the central region is processed into an integral image to obtain a two-dimensional integral image; the two-dimensional integral image is processed into a sliding window to obtain the detection window and value of each detection point; the CFAR value of each detection point is calculated based on the CFAR threshold and the detection window and value; the binary image of the central region is obtained by comparing each detection point of the SAR two-dimensional amplitude image with the CFAR value of each detection point.

[0143] In this embodiment, the sliding window processing of the two-dimensional integral image to obtain the detection window and value for each detection point includes:

[0144] Please refer to Figure 3 The black area represents the point to be tested, the yellow area represents the outer window of that point, and the red area represents the protective window of the point. For example... Figure 4 The diagram shown is a schematic of the detection points. First, a sliding window process is applied to the two-dimensional integral graph to obtain the outer window sum for each detection point:

[0145] sum1 i =rb i +lux-ru i -lb i ;

[0146] Among them, rb i Let lu be the value of the lower right point of the outer window of the i-th detection point in the two-dimensional integral image. i Let ru be the value of the top-left point of the outer window of the i-th detection point in the two-dimensional integral graph. i Let lb be the value of the upper right point of the outer window of the i-th detection point in the two-dimensional integral graph. i The sum1 is the value of the lower left point of the outer window of the i-th detection point in the two-dimensional integral graph; i Let the sum of the outer windows of the i-th detection point be ;

[0147] Then, a sliding window process is performed on the two-dimensional integral image to obtain the inner window sum for each detection point;

[0148] The detection window and value for each detection point are:

[0149] sum i =sum1 i -sum2 i ;

[0150] Where, sum i Let sum1 be the sum of the detection window values ​​at the i-th detection point, sum2 be the sum of the outer window values ​​at the i-th detection point, and sum3 be the sum of the outer window values ​​at the i-th detection point. i Let be the sum of the inner windows of the i-th detection point.

[0151] In this embodiment, calculating the CFAR value for each detection point based on the CFAR threshold and the detection window sum includes:

[0152] cfar_yi =sum i ×cen_pfa

[0153] Where, sum i Here, cen_pfa represents the detection window and value at the i-th detection point in the central region, and cen_pfa represents the CFAR threshold at the i-th detection point in the central region; cfar_y i is the cfar value of the i-th detection point in the central region.

[0154] CFA detection is performed on the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions to obtain binary maps of the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions, respectively.

[0155] like Figure 2 As shown in the image, the yellow areas represent the four corner regions. CFAR detection is performed on each of these regions, resulting in binary images of the corner regions. CFAR detection of the corner boundary regions includes:

[0156] cfar_L=((Det_L+Pro_L)2+1)×((Det_L+Pro_L)*2+1);

[0157] Where Det_L is the length of the detection window and Pro_L is the length of the protection window;

[0158] Calculate the CFA detection values ​​of the four corner points of the four-corner boundary region based on the CFA value of each detection point in the central region;

[0159] lu_th=sum [0] *cen_pfa;

[0160] ru_th=sum [N-2×(Det_l+Pro_l)-1] ×cen_pfa;

[0161] lb_th = sum [M-2×(Det_l+Pro_l-1))*N] ×cen_pfa;

[0162] rb_th = sum [(M-2×(Det_l+Pro_l )-1)*N+N-2×(Det_l+Pro_l)-1]×cen_pfa;

[0163] Where, lu_th is the CFAR detection value of the top-left corner, ru_th is the CFAR detection value of the top-right corner, lb_th is the CFAR detection value of the bottom-left corner, and rb_th is the CFAR detection value of the bottom-right corner; M represents the length of the range direction of the two-dimensional range-Doppler SAR image, and N represents the length of the azimuth direction of the two-dimensional range-Doppler SAR image; Det_L is the length of the detection window, Pro_L is the length of the protection window; sum [0]For the outer window of the 0th detection point, sum [N-2×(Det_l+Pro_l)-1] The sum of the outer window values ​​at the (N-2×(det_l+pro_l)-1)th detection point is given by sum. [(M-2×(Det_l+Pro_l-1))*N] Let sum be the sum of the outer windows of the (M-2×(Det_l+Pro_l)-1)*Nth detection point. [(M-2×(Det_l+Pro_l)-1)*N+N-2×(Det_l+Pro_l)-1] The sum of the outer window of the (M-2×(Det_l+Pro_l)-1)*N+N-2×(Det_l+Pro_l)-1 detection point.

[0164] In this embodiment, CFAR detection of the four corner boundary regions further includes:

[0165] Each detection point in the upper left corner is compared with the cfar detection value lu_th of the upper left corner point. If the detection point is greater than lu_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, and the binary image of the upper left corner can be obtained.

[0166] Each detection point in the upper right corner is compared with the cfar detection value ru_th of the upper right corner point. If the detection point is greater than re_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, and the binary image of the upper right corner can be obtained.

[0167] Each detection point in the lower left corner is compared with the cfar detection value lb_th of the lower left corner point. If the detection point is greater than lb_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, and the binary image of the lower left corner can be obtained.

[0168] Each detection point in the lower right corner is compared with the cfar detection value rb_th of the lower right corner point. If the detection point is greater than rb_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the lower right corner.

[0169] In this embodiment, performing CFAR detection on the left and right boundary regions includes:

[0170] like Figure 5 As shown, the detection window for the left boundary region uses the right half-window CFAR detection, while the detection window for the right boundary region uses the left half-window CFAR detection.

[0171] The sliding window size used for CFA detection of the left and right boundary regions is:

[0172] dpl_win_S=(Det_L+Pro_L+1)×((Det_L+Pro_L)×2+1);

[0173] Where dpl_win_S is the size of the sliding window used for CFA detection of the left and right boundary regions. Det_L is the length of the detection window, and Pro_L is the length of the protection window;

[0174] Compare all detection points within the left and right boundary regions with the corresponding CFAR detection values. If a detection point is larger than the CFAR detection value, set the value to 1; otherwise, set it to 0. This will give you the binary map of the upper region.

[0175] In this embodiment, performing CFAR detection on the left boundary region specifically includes:

[0176] The sum of the outer windows of the b-th detection point in the first row of the left boundary region is:

[0177] huge_win_sum[b]=huge_win_sum[b-1]-huge_sum_col[b]+huge_sum_col[b+Pro_L+Det_L];

[0178] Where huge_win_sum[b] is the sum of the outer windows of the b-th sliding window block in the first row, huge_win_sum[b-1] is the sum of the outer windows of the (b-1)-th sliding window block in the first row, huge_sum_col[b] is the sum of the outer windows of the detection points in the b-th column of the first row, and huge_sum_col[b+Pro_L+Det_L] is the sum of the outer windows of the detection points in the b+pro_L+Det_L column of the first row;

[0179] The sum of the inner windows at the b-th detection point in the first row of the left boundary region is:

[0180] pro_win_sum[b]=pro_win_sum[b-1]+small_sum_col[b+Pro_L]-sumall_sum_col[b];

[0181] Where pro_win_sum[b] is the inner window sum of the b-th detection point in the first row, pro_win_sum[b-1] is the inner window sum of the (b-1)-th detection point in the first row, small_sum_col[b+Pro_L] is the inner window sum of the b+Pro_L-th column in the first row, and sumall_sum_col[b] is the inner window sum of the b-th column in the first row;

[0182] The detection window sum value corresponding to the b-th detection point in the first row within the left boundary region is:

[0183] det_win_sum[b]=huge_win_sum[b]-pro_win_sum[b];

[0184] Where huge_win_sum[b] is the outer window sum of the b-th detection point in the first row; pro_win_sum[b] is the inner window sum of the b-th detection point in the first row; and det_win_sum[b] is the detection window sum value corresponding to the b-th detection point in the first row;

[0185] The sum of the outer windows of the detection points in the q-th row and k-th column of the left boundary region, excluding the first row, is:

[0186] huge_sum_col[k]=huge_sum_col[k-1]-cfar_sb[(q-Pro_L-Dis_L-1)×N+k]+cfar_sb[(q+Pro_L+Dis_L)×N+k];

[0187] huge_sum_col[k] is the sum of the outer windows in the q-th row and k-th column excluding the first row, huge_sum_col[k-1] is the sum of the outer windows in the q-th row and (k-1)-th column excluding the first row, cfar_sb[(q-Pro_L-Dis_L-1)×N+k] is the pixel value of the (q-Pro_L-Dis_L-1)×N+k-th point on the SAR image of the left boundary region, where N represents the length of the azimuth direction of the two-dimensional distance Doppler SAR image; cfar_sb(q+Pro_L+Dis_L)×N+k is the pixel value of the (q-Pro_L-Dis_L-1)×N+k-th point on the SAR image of the left boundary region.

[0188] The sum of the inner windows of the detection points in the q-th row and k-th column of the left boundary region, excluding the first row, is:

[0189] small_sum_col[k]=small_sum_col[k-1]-cfar_sb[(q-Pro_L-1)×N+k]+cfar_sb[(q+Pro_L)×N+k];

[0190] Where, small_sum_col[k-1] is the sum of the inner windows in the q-th row and k-th column excluding the first row, small_sum_col[k] is the sum of the inner windows in the q-th row and k-th column excluding the first row, cfar_sb is the length of the azimuth direction of the SAR image N, representing the two-dimensional distance Doppler SAR image; cfar_sb[(q-Pro_L-1)×N+k] is the (q-Pro_L-1)×N+k pixel on the SAR image of the left boundary region; cfar_sb[(q+Pro_L)×N+k] is the [(q+Pro_L)×N+k] pixel on the SAR image of the left boundary region;

[0191] The detection window sum value corresponding to the detection point in the q-th row and k-th column of the left boundary region (excluding the first row) is:

[0192] det_win_sum[k]=huge_sum_col[k]-small_sum_col[k];

[0193] The half-window detection area of ​​the left boundary region is:

[0194] S_det_dpl=((det_L+Pro_L)×2+1)×(det_L+Pro_L+1)-(Pro_l+1)×Pro_L×2+1);

[0195] The CFA detection values ​​for the detection points in the left half of the region are:

[0196]

[0197] Among them, dpl_hf_pfa b is the CFAR detection value of the b-th detection point within the left boundary region; cfar_sb is the detection point of the SAR image within the left boundary region; S_det_dpl is the area of ​​the left boundary region.

[0198] Compare all detection points within the left boundary region with the corresponding CFAR detection value. If a detection point is larger than the CFAR detection value, set the value to 1; otherwise, set it to 0. This will give you the binary map of the upper region.

[0199] like Figure 5 As shown, the upper boundary region detection window uses the lower half-window CFAR detection, and the lower boundary region detection window uses the upper half-window CFAR detection. CFAR detection for both the upper and lower boundary regions includes:

[0200] The upper boundary region detection window uses the lower half-window CFAR detection, while the lower boundary region detection window uses the upper half-window CFAR detection.

[0201] The sizes of the sliding windows in the upper and lower boundary areas are:

[0202] cfar_dis_win=(Det_L+Pro_L+1)×((Det_L+Pro_L)×2+1);

[0203] Where Det_L is the detection window length and Pro_L is the protection window length;

[0204] Adding the sums of each column within the first sliding window of the first row will give you the sum of the outer window of the first sliding window in the first row: sum_win_dis1;

[0205] The sum of the outer windows of the first sliding window in the first row is used to obtain the sum of the outer windows of the second sliding window; the sum of the outer windows of the m-th sliding window in the first row is used to obtain the sum of the outer windows of the (m+1)-th sliding window in the first row.

[0206] The sum of the outer windows of the (m+1)th sliding window in the first row is:

[0207] sum_win_dis m+1 =sum_win_dis m +dis_sum_col[m+Det_L+Pro_L+1]-dis_sum_col[m-Det_L+Pro_L];

[0208] Among them, sum_win_dis m+1 The sum of the outer windows of the (m+1)th sliding window in the first row of the upper boundary region, sum_win_dis m The sum of the outer windows of the m-th sliding window in the first row of the upper boundary region is given by dis_sum_col[m+Det_L+Pro_L+1], where m+Det_L+Pro_L+1 is the sum of the outer windows of the m-th sliding window in the first row of the upper boundary region.

[0209] The sum of the outer window in column 1, dis_sum_col[m-Det_L+Pro_L] is the sum of the outer window in column m-Det_L+Pro_L within the first row of the sliding window block in the upper boundary region;

[0210] Adding the sums of the small windows in each column of the first sliding window in the first row yields the sum of the protection windows for the first sliding window in the first row. Based on the sum of the protection windows of the first sliding window in the first row, we obtain the sum of the protection windows of the second sliding window in the first row; based on the sum of the protection windows of the second sliding window in the first row, we obtain the sum of the protection windows of the third sliding window in the first row; based on the sum of the protection windows of the m-th sliding window in the first row, we obtain the sum of the (m+1)-th protection windows in the first row; the sum of the protection windows corresponding to the m-th detection point in the first row is: sum_small_win_dis m ;

[0211] The sum of the outer windows of the detection points in the r-th row and t-th column, excluding the first row of the upper boundary region, is:

[0212] dis_sum_col[t]=dis_sum_col[t-1]-cfar_sm[(r-1)×N+t]+cfar_sm[(r+Det_L+Pro_L)*N+t];

[0213] Where, dis_sum_col[t] is the sum of the outer windows in the r-th row and t-th column excluding the first row of the upper boundary region, dis_sum_col[t-1] is the sum of the outer windows in the r-th row and t-1-th column excluding the first row of the upper boundary region, cfar_sm[(r-1)×N+t] is the pixel value of the (r-1)×N+t-th point on the SAR image of the upper boundary region, and cfar_sm[(r+Det_L+Pro_L)*N+t] is the pixel value of the (r+Det_L+Pro_L)×N+t-th point on the SAR image of the upper boundary region;

[0214] The sum of the inner windows of the detection points in the r-th row and t-th column, excluding the first row of the upper boundary region, is:

[0215] dis_sum_small_col[t]=dis_sum_small_col[t-1]-cfar_sm[(r-1)×N+t]+cfar_sm[(r+Det_L+Pro_L)*N+t];

[0216] Where, dis_sum_small_col[t] is the inner window sum of the r-th row and t-th column excluding the first row, dis_sum_small_col[t-1] is the inner window sum of the r-th row and t-1-th column excluding the first row, cfar_sm[(r-1)×N+t] is the pixel value of the (r-1)×N+t-th point on the SAR image of the upper boundary region, and cfar_sm[(+Det_L+Pro_L)*N+t] is the pixel value of the (r+Det_L+Pro_L)×N+t-th point on the SAR image of the upper boundary region;

[0217] The detection window and value sum_det_dis of the detection points in the first row of the upper boundary region m for:

[0218] sum_det_dis m =sum_win_dis m -sum_small_win_dis m ;

[0219] The detection window and value sum_det_dis of the detection points excluding the first row in the upper boundary region t for:

[0220] sum_det_dis t =dis_sum_small_col[t]-sum_det_dis t ;

[0221] The half-window detection area of ​​the upper boundary region is:

[0222] S_det_dis=((det_L+Pro_L)×2+1)×(det_L+Pro_L+1-(Pro_L+1)×(Pro_L×2+1);

[0223] The CFA detection value of the detection points in the upper boundary region is:

[0224]

[0225] Where cfar_sm is the pixel value of the m-th point on the SAR image of the upper boundary region; dis_hf_pfa m S_det_dis is the CFA detection value of the m-th detection point in the upper boundary region; S_det_dis is the half-window detection area of ​​the upper boundary region.

[0226] For any detection point in the upper or lower boundary region, if the amplitude value of the detection point is greater than its corresponding CFA detection value, then the detection point is set to 1; otherwise, it is set to 0.

[0227] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A fast two-dimensional CFAR detection method for SAR imaging recognition, characterized in that, The method includes: The two-dimensional distance Doppler SAR image is converted into a two-dimensional SAR amplitude image. The two-dimensional SAR amplitude image is then divided into four regions: the central region, the four corner boundary regions, the left and right boundary regions, and the upper and lower boundary regions. The Doppler value range of the central region is obtained based on the distance value range of the central region; the area of ​​the annular region of the detection window in the central region is calculated based on the Doppler value range of the central region; and the CFAR threshold is calculated based on the area of ​​the annular region of the detection window. The SAR two-dimensional amplitude image of the central region is processed into an integral image to obtain a two-dimensional integral image; the two-dimensional integral image is processed into a sliding window to obtain the detection window and value of each detection point; the CFAR value of each detection point is calculated based on the CFAR threshold and the detection window and value; the binary image of the central region is obtained by comparing each detection point of the SAR two-dimensional amplitude image with the CFAR value of each detection point. CFA detection is performed on the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions to obtain binary maps of the four corner boundary regions, left and right boundary regions, and top and bottom boundary regions, respectively.

2. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 1, characterized in that, The Doppler range of the central region, obtained based on the distance range of the central region, includes: First, obtain the range of distance values ​​for the central region: [(Pro_Dis_L+Dest_Dis_L-1),M-(Pro_Dis_L+Dest_Dis_L-1)]; Where Pro_Dis_L is the half-window length of the protection window in the distance dimension of the central region, M is the distance dimension length, and Dest_Dis_L is the half-window length of the detection window in the distance dimension of the central region. The range of Doppler values ​​in the central region is: [(Pro_Dpl_L+Dest_Dpl_L-1),N-(Pro_Dpl_L+Dest_Dpl_L-1)]; Where Pro_Dpl_L is the half-window length of the Doppler protection window, N is the Doppler length, and Dest_Dpl_L is the half-window length of the Doppler detection window.

3. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 2, characterized in that, Methods for calculating the CFAR threshold include: First, calculate the area of ​​the annular region of the detection window: cen_win_S=((Det_L+Pro_L)×2+1)×((Det_L+Pro_L)×2+1)-(Pro_L×2+1)×(Pro_L×2+1); Where cen_win_S is the area of ​​the annular region of the detection window, Det_L is the length of the Doppler detection window, and Pro_L is the length of the Doppler protection window; Then, the cfar threshold is calculated based on the area of ​​the annular region of the detection window; Wherein, CFAR_PFA is the CFAR constant false alarm factor, cen_win_S is the area of ​​the annular region of the detection window, and cen_pfa is the CFAR threshold.

4. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 3, characterized in that, Applying a sliding window technique to the two-dimensional integral image yields the detection window and value for each detection point, including: Perform sliding window processing on the two-dimensional integral image and calculate the sum of the outer windows for each detection point: sum1 i =rb i +lu i -ru i -lb i 4 Among them, rb i Let lu be the value of the lower right point of the outer window of the i-th detection point in the two-dimensional integral image. i Let ru be the value of the top-left point of the outer window of the i-th detection point in the two-dimensional integral graph. i Let lb be the value of the upper right point of the outer window of the i-th detection point in the two-dimensional integral graph. i The sum1 is the value of the lower left point of the outer window of the i-th detection point in the two-dimensional integral graph; i Let the sum of the outer windows of the i-th detection point be ; The two-dimensional integral image is subjected to sliding window processing to obtain the inner window sum of each detection point; The detection window and value for each detection point are: sum i =sum1 i -sum2 i ; Where, sum i Let sum1 be the sum of the detection window values ​​at the i-th detection point, sum2 be the sum of the outer window values ​​at the i-th detection point, and sum3 be the sum of the outer window values ​​at the i-th detection point. i Let be the sum of the inner windows of the i-th detection point.

5. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 4, characterized in that, The calculation of the CFAR value for each detection point based on the CFAR threshold and detection window sum includes: cfar_y i =sum i ×cen_pfa; Where, sum i Here, cen_pfa represents the detection window and value at the i-th detection point in the central region, and cen_pfa represents the CFAR threshold at the i-th detection point in the central region; cfar_y i is the cfar value of the i-th detection point in the central region.

6. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 5, characterized in that, CFAR detection of the four corner boundary regions includes: The size of the four corner boundary regions is: cfar_L=((Det_L+Pro_L)*2+1)×((Det_L+Pro_L)*2+1); Where Det_L is the length of the detection window, Pro_L is the length of the protection window, and cfar_L is the size of the four corner boundary regions; Calculate the CFA detection values ​​of the four corner points of the four-corner boundary region based on the CFA value of each detection point in the central region; lu_th=sum [0 x*cen_pfa: ru_th=sum [N-2×(Det_l+Pro_l)-1] ×cen_pfa; lb_th=sum [(M-2×(Det_l+Pro_l-1))*N] ×cen_pfa; rb_th=sum [(M-2×(Det_l+Pro_l)-1)*N+N-2×(Det_l+Pro_l)-1] ×cen_pfa; Where, lu_th is the CFAR detection value of the top-left corner, ru_th is the CFAR detection value of the top-right corner, lb_th is the CFAR detection value of the bottom-left corner, and rb_th is the CFAR detection value of the bottom-right corner; M represents the length of the range direction of the two-dimensional range-Doppler SAR image, and N represents the length of the azimuth direction of the two-dimensional range-Doppler SAR image; Det_L is the length of the detection window, Pro_L is the length of the protection window; sum [0] For the outer window of the 0th detection point, sum [N-2×(Det_l+Pro_l)-1] The sum of the outer window values ​​at the (N-2×(det_l+pro_l)-1)th detection point is given by sum. [(M-2×(Det_l+Pro_l-1))*N] Let sum be the sum of the outer windows of the (M-2×(Det_l+Pro_l)-1)*Nth detection point. [(M-2×(Det_l+Pro_l)-1)*N+N-2×(Det_l+Pro_l)-1] The sum of the outer window of the (M-2×(Det_l+Pro_l)-1)*N+N-2×(Det_l+Pro_l)-1 detection point.

7. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 6, characterized in that, CFAR detection of the four corner boundary regions also includes: Each detection point in the upper left corner is compared with the cfar detection value lu_th of the upper left corner point. If the detection point is greater than lu_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the upper left corner. Each detection point in the upper right corner is compared with the cfar detection value ru_th of the upper right corner point. If the detection point is greater than ru_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the upper right corner. Each detection point in the lower left corner is compared with the cfar detection value lb_th of the lower left corner point. If the detection point is greater than lb_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the lower left corner. Each detection point in the lower right corner is compared with the cfar detection value rb_th of the lower right corner point. If the detection point is greater than rb_th, the pixel value of the detection point is set to 1, otherwise it is set to 0, thus obtaining the binary image of the lower right corner.

8. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 7, characterized in that, CFAR detection of the left and right boundary regions includes: The left boundary region detection window uses the right half-window CFAR detection, while the right boundary region detection window uses the left half-window CFAR detection. The sliding window size used for CFA detection of the left and right boundary regions is: dpl_win_S=(Det_L+Pro_L+1)×((Det_L+Pro_L)×2+1); Where dpl_win_S is the size of the sliding window used for CFA detection of the left and right boundary regions, Det_L is the length of the detection window, and Pro_L is the length of the protection window; Compare all detection points within the left and right boundary regions with the corresponding CFAR detection values. If a detection point is larger than the CFAR detection value, set the value to 1; otherwise, set it to 0. This will give you the binary map of the upper region.

9. The rapid two-dimensional CFAR detection method for SAR imaging recognition according to claim 8, characterized in that, The specific steps for performing CFAR detection on the left boundary region include: The sum of the outer windows of the b-th detection point in the first row of the left boundary region is: huge_win_sum[b]=huge_win_sum[b-1]-huge_sum_col[b]+huge_sum_col[b+Pro_L+Det_L]; Where huge_win_sum[b] is the sum of the outer windows of the b-th sliding window block in the first row, huge_win_sum[b-1] is the sum of the outer windows of the (b-1)-th sliding window block in the first row, huge_sum_col[b] is the sum of the outer windows of the detection points in the b-th column of the first row, and huge_sum_col[b+Pro_L+Det_L] is the sum of the outer windows of the detection points in the b+pro_L+Det_L column of the first row; The sum of the inner windows at the b-th detection point in the first row of the left boundary region is: pro_win_sum[b]=pro_win_sum[b-1]+small_sum_col[b+Pro_L]-sumall_sum_col[b]; Where pro_win_sum[b] is the inner window sum of the b-th detection point in the first row, pro_win_sum[b-1] is the inner window sum of the (b-1)-th detection point in the first row, small_sum_col[b+Pro_L] is the inner window sum of the b+Pro_L-th column in the first row, and sumall_sum_col[b] is the inner window sum of the b-th column in the first row; The sum of the detection window values ​​corresponding to the b-th detection point in the first row within the left boundary region is: det_win_sum[b]=huge_win_sum[b]-pro_win_sum[b]; Where huge_win_sum[b] is the outer window sum of the b-th detection point in the first row; pro_win_sum[b] is the inner window sum of the b-th detection point in the first row; and det_win_sum[b] is the detection window sum value corresponding to the b-th detection point in the first row; The sum of the outer windows of the detection points in the q-th row and k-th column of the left boundary region, excluding the first row, is: huge_sum_col[k]=huge_sum_col[k-1]-cfar_sb[(q-Pro_L-Dis_L-1)×N+k]+cfar_sb[(q+Pro_L+Dis_L)×N+k]; huge_sum_col[k] is the sum of the outer windows in the q-th row and k-th column excluding the first row, huge_sum_col[k-1] is the sum of the outer windows in the q-th row and (k-1)-th column excluding the first row, cfar_sb[(q-Pro_L-Dis_L-1)×N+k] is the pixel value of the (q-Pro_L-Dis_L-1)×N+k-th point on the SAR image of the left boundary region, where N represents the length of the azimuth direction of the two-dimensional distance Doppler SAR image; cfar_sb(q+Pro_L+Dis_L)×N+k is the pixel value of the (q-Pro_L-Dis_L-1)×N+k-th point on the SAR image of the left boundary region. The sum of the inner windows of the detection points in the q-th row and k-th column of the left boundary region, excluding the first row, is: small_sum_col[k]=small_sum_col[k-1]-cfar_sb[(q-Pro_L-1)×N+k]+cfar_sb[(q+Pro_L)×N+k]; Where, small_sum_col[k-1] is the sum of the inner windows in the q-th row and k-th column excluding the first row, small_sum_col[k] is the sum of the inner windows in the q-th row and k-th column excluding the first row, cfar_sb is the length of the azimuth direction of the SAR image N, representing the two-dimensional distance Doppler SAR image; cfar_sb[(q-Pro_L-1)×N+k] is the (q-Pro_L-1)×N+k pixel on the SAR image of the left boundary region; cfar_sb[(q+Pro_L)×N+k] is the [(q+Pro_L)×N+k] pixel on the SAR image of the left boundary region; The detection window sum value corresponding to the detection point in the q-th row and k-th column of the left boundary region (excluding the first row) is: det_win_sum[k]=huge_sum_col[k]-small_sum_col[k]; The half-window detection area of ​​the left boundary region is: S_det_dpl=((det_L+Pro_L)×2+1)×(det_L+Pro_L+1)-(Pro_L+1)×(Pro_L×2+1); The CFA detection values ​​for the detection points in the left half of the region are: Where dpl_hf_pfab is the CFAR detection value of the b-th detection point within the left boundary region; cfar_sb is the detection point of the SAR image within the left boundary region; and S_det_dpl is the area of ​​the left boundary region. Compare all detection points within the left boundary region with the corresponding CFAR detection value. If a detection point is larger than the CFAR detection value, set the value to 1; otherwise, set it to 0. This will give you the binary map of the upper region.

10. The SAR imaging recognition method for rapid two-dimensional CFAR detection according to claim 9, characterized in that, CFAR detection of the upper and lower boundary regions includes: The upper boundary region detection window uses the lower half-window CFAR detection, while the lower boundary region detection window uses the upper half-window CFAR detection. The sizes of the sliding windows in the upper and lower boundary areas are: cfar_dis_win=(Det_L+Pro_L+1)×((Det_L+Pro_L)×2+1); Where Det_L is the detection window length and Pro_L is the protection window length; Adding the sums of each column within the first sliding window of the first row will give you the sum of the outer window of the first sliding window in the first row: sum_win_dis1; The sum of the outer windows of the first sliding window in the first row is used to obtain the sum of the outer windows of the second sliding window; the sum of the outer windows of the m-th sliding window in the first row is used to obtain the sum of the outer windows of the (m+1)-th sliding window in the first row. The sum of the outer windows of the (m+1)th sliding window in the first row is: sum_win_dis m+1 sum_win_dis m +this_sum_col[m+L+Pro_L+1]-this_sum_col[m-L+Pro_L] Among them, sum_win_dis m+1 The sum of the outer windows of the (m+1)th sliding window in the first row of the upper boundary region, sum_win_dis m For the m-th sliding window in the first row of the upper boundary region, dis_sum_col[m+Det_L+Pro_L+1] is the sum of the outer windows in the m+Det_L+Pro_L+1-th column of the sliding window block in the first row of the upper boundary region, and dis_sum_col[m-Det_L+Pro_L] is the sum of the outer windows in the m-Det_L+Pro_L-th column of the sliding window block in the first row of the upper boundary region. The sum of the protection windows of the first sliding window in the first row is used to obtain the sum of the protection windows of the second sliding window in the first row; the sum of the protection windows of the second sliding window in the first row is used to obtain the sum of the protection windows of the third sliding window in the first row; the sum of the protection windows of the m-th sliding window in the first row is used to obtain the sum of the (m+1)-th protection windows in the first row; the sum of the protection windows corresponding to the m-th detection point in the first row is: sum_small_win_dis m ; The sum of the outer windows of the detection points in the r-th row and t-th column, excluding the first row of the upper boundary region, is: dis_sum_col[t]=dis_sum_col[t-1]-cfar_sm[(r-1)×N+t]+cfar_sm[(r+Det_L+Pro_L)*N+t]; Where, dis_sum_col[t] is the sum of the outer windows in the r-th row and t-th column excluding the first row of the upper boundary region, dis_sum_col[t-1] is the sum of the outer windows in the r-th row and t-1-th column excluding the first row of the upper boundary region, cfar_sm[(r-1)×N+t] is the pixel value of the (r-1)×N+t-th point on the SAR image of the upper boundary region, and cfar_sm[(r+Det_L+Pro_L)*N+t] is the pixel value of the (r+Det_L+Pro_L)×N+t-th point on the SAR image of the upper boundary region; The sum of the inner windows of the detection points in the r-th row and t-th column, excluding the first row of the upper boundary region, is: dis_sum_small_col[t]=dis_sum_small_col[t-1]-cfar_sm[(r-1)×N+t]+cfar_sm[(r+Det_L+Pro_L)*N+t]; Where, dis_sum_small_col[t] is the inner window sum of the r-th row and t-th column excluding the first row, dis_sum_small_col[t-1] is the inner window sum of the r-th row and t-1-th column excluding the first row, cfar_sm[(r-1)×N+t] is the pixel value of the (r-1)×N+t-th point on the SAR image of the upper boundary region, and cfar_sm[(r+Det_L+Pro_L)*N+t] is the pixel value of the (r+Det_L+Pro_L)×N+t-th point on the SAR image of the upper boundary region; The detection window and value sum_det_dis of the detection points in the first row of the upper boundary region m for: sum_it_this m sum_win_dis m -sum_small_win_dis m 100. The detection window and value sum_det_dis of the detection points excluding the first row in the upper boundary region t for: sum_the_dis t =dis_sum_small_col[t]-sum_det_dis t ; The half-window detection area of ​​the upper boundary region is: S_det_dis=((det_L+Pro_L)×2+1)×(det_L+Pro_L+1)-(Pro_L+1)×(Pro_L×2+1); The CFA detection value of the detection points in the upper boundary region is: Where cfar_sm is the pixel value of the m-th point on the SAR image of the upper boundary region; dis_hf_pfa m S_det_dis is the CFA detection value of the m-th detection point in the upper boundary region; S_det_dis is the half-window detection area of ​​the upper boundary region. For any detection point in the upper or lower boundary region, if the amplitude value of the detection point is greater than its corresponding CFA detection value, then the detection point is set to 1; otherwise, it is set to 0.

Images