High-resolution SAR satellite image line segment extraction method

Abstract

The invention relates to a high-resolution SAR satellite image line segment extraction method, and the method comprises the steps: processing an image into an amplitude image, and obtaining a seed point set based on constant false alarm rate detection, non-maximum suppression and diagonal point response value thresholding; obtaining a candidate growing point set based on constant false alarm rate detection, non-maximum suppression and an adaptive amplitude threshold of the amplitude image; constructing an annular neighborhood window by taking each point in the seed point and candidate growing point set as a center, and obtaining a main direction of each point through weighted principal component analysis based on a pixel amplitude value of a pixel point in the annular neighborhood window of each point; and for each seed point, based on the candidate growing point set, determining a first growing point by using spatial distance and direction consistency constraints, and determining a current second growing point by using spatial distance, direction consistency and fitting direction vector change constraints, so that high-precision line segment extraction can be performed on linear features formed by discrete strong scattering points in the image.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of high-resolution SAR satellite image information extraction, and particularly relates to a high-resolution SAR satellite image line segment extraction method. BACKGROUND

[0002] With the development of high-resolution synthetic aperture radar (SAR) satellite technology, the image data provided by high-resolution SAR satellites has high spatial resolution. The high-resolution SAR satellite image contains a large number of linear features composed of discrete strong scattering points, which are mainly derived from building facades, bridges and other structure strong reflections. The line segments obtained by extracting the linear features have important application value in the fields of building structure analysis and three-dimensional modeling.

[0003] Existing image line segment extraction methods mainly include gradient-based edge detection and Hough transform based on transform domain. These existing methods, mostly derived from the field of computer vision, can only extract the linear features of the face-shaped target boundary line in the SAR satellite image with high precision, and cannot extract the linear features composed of discrete strong scattering points in the high-resolution SAR satellite image with high precision. SUMMARY

[0004] The present application provides a high-resolution SAR satellite image line segment extraction method, which can realize high-precision line segment extraction of linear features composed of discrete strong scattering points in high-resolution SAR satellite images.

[0005] In a first aspect, the embodiments of the present application provide a high-resolution SAR satellite image line segment extraction method, comprising: obtaining a high-resolution SAR satellite image, processing the image into an amplitude image, obtaining a seed point set based on constant false alarm rate detection, non-maximum suppression and diagonal point response thresholding, and obtaining a candidate growth point set based on non-maximum suppression and adaptive amplitude threshold of the amplitude image; constructing a ring neighborhood window with each point in the seed point set and the candidate growth point set as the center, obtaining the main direction of each seed point in the seed point set and the main direction of each candidate growth point in the candidate growth point set by weighted principal component analysis based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point; and For each sub-point, based on the main direction of the seed point and the main direction of the candidate growth point, a first growth point is determined from the candidate growth point set by using spatial distance and direction consistency constraints, and an initial fitting line segment is generated; a step growth step: step growth is performed in the positive and negative directions of the initial fitting line segment or the current fitting line segment to obtain a current step growth line segment; a second growth point determination step: based on the end point of the current step growth line segment, a current second growth point is determined from the candidate growth point set by using spatial distance, direction consistency and fitting direction vector change constraints, and a current fitting line segment is generated; the step growth step and the current second growth point determination step are repeated until a new second growth point cannot be found.

[0006] In a second aspect, an embodiment of the present application provides another high-resolution SAR satellite image line segment extraction method, comprising: A high-resolution SAR satellite image is acquired, and the image is processed into an amplitude image; a seed point set is obtained based on constant false alarm rate detection, non-maximum suppression and diagonal point response value thresholding; a candidate growth point set is obtained based on non-maximum suppression and adaptive amplitude thresholding of the amplitude image; A ring neighborhood window is constructed with each point in the seed point set and the candidate growth point set as the center, and the main direction of each sub-point in the seed point set and the main direction of each candidate growth point in the candidate growth point set are obtained by weighted principal component analysis based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point; For each sub-point, based on the main direction of the seed point and the main direction of the candidate growth point, a first growth point is determined from the candidate growth point set by using spatial distance and direction consistency constraints; based on the first growth point, a second growth point is determined from the candidate growth point set by using spatial distance, direction consistency and fitting direction vector change constraints, and a current fitting line segment is generated; a new second growth point determination step: based on the current second growth point and the current fitting line segment, a new second growth point is determined from the candidate growth point set by using spatial distance, direction consistency and fitting direction vector change constraints, and a new fitting line segment is generated; the new second growth point determination step is repeated until a new second growth point cannot be found.

[0007] In a third aspect, an embodiment of the present application provides a high-resolution SAR satellite image line segment extraction system, comprising: A growth point determination module is configured to: acquire a high-resolution SAR satellite image, process the image into an amplitude image, obtain a seed point set based on constant false alarm rate detection, non-maximum suppression and diagonal point response value thresholding, and obtain a candidate growth point set based on non-maximum suppression and adaptive amplitude thresholding of the amplitude image; The main direction determination module of the growth point is configured to: construct a ring neighborhood window with each point in the seed point set and the candidate growth point set as the center, obtain the main direction of each seed point in the seed point set and the main direction of each candidate growth point in the candidate growth point set based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point through weighted principal component analysis, and determine the first growth point from the candidate growth point set based on the main direction of the seed point and the main direction of the candidate growth point by using the spatial distance and direction consistency constraint. The line segment growth module is configured to: determine the first growth point from the candidate growth point set based on the main direction of the seed point and the main direction of the candidate growth point by using the spatial distance and direction consistency constraint to generate the initial fitting line segment for each seed point; perform step growth in the positive and negative directions of the initial fitting line segment or the current fitting line segment to obtain the current step growth line segment; determine the current second growth point from the candidate growth point set based on the end point of the current step growth line segment by using the spatial distance, direction consistency and fitting direction vector change constraint to generate the current fitting line segment; and repeat the step growth step and the current second growth point determination step until a new second growth point cannot be found.

[0008] In a fourth aspect, an embodiment of the present application provides a high-resolution SAR satellite image line segment extraction system, comprising: The growth point determination module is configured to: obtain a high-resolution SAR satellite image, process the image into an amplitude image, obtain a seed point set based on constant false alarm rate detection, non-maximum suppression and diagonal point response thresholding, and obtain a candidate growth point set based on non-maximum suppression and an adaptive amplitude threshold of the amplitude image. The main direction determination module of the growth point is configured to: construct a ring neighborhood window with each point in the seed point set and the candidate growth point set as the center, obtain the main direction of each seed point in the seed point set and the main direction of each candidate growth point in the candidate growth point set based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point through weighted principal component analysis, and determine the first growth point from the candidate growth point set based on the main direction of the seed point and the main direction of the candidate growth point by using the spatial distance and direction consistency constraint. The line segment growth module is configured to: determine the first growth point from the candidate growth point set based on the main direction of the seed point and the main direction of the candidate growth point by using the spatial distance and direction consistency constraint to generate the initial fitting line segment for each seed point; determine the second growth point from the candidate growth point set based on the first growth point by using the spatial distance, direction consistency and fitting direction vector change constraint to generate the current fitting line segment; determine the new second growth point from the candidate growth point set based on the current second growth point and the current fitting line segment by using the spatial distance, direction consistency and fitting direction vector change constraint to generate the new fitting line segment; and repeat the new second growth point determination step until a new second growth point cannot be found.

[0009] In a fifth aspect, an electronic device is provided, which includes a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the high-resolution SAR satellite image line segment extraction method provided in the first aspect or the second aspect when executing the computer program.

[0010] In a sixth aspect, a computer readable storage medium is provided, which stores a computer program, and the computer program implements the steps of the high-resolution SAR satellite image line segment extraction method provided in the first aspect or the second aspect when executed by a processor.

[0011] In a seventh aspect, a computer program product is provided, which includes a computer program, and the computer program implements the steps of the high-resolution SAR satellite image line segment extraction method provided in the first aspect or the second aspect when executed by a processor.

[0012] The method provided in the present application, in the first aspect, uses constant false alarm rate (CFAR) detection to find candidate strong scattering points from the image; candidate growth points are obtained by non-maximum suppression (NMS) in the candidate strong scattering points; further, noise points in the candidate strong scattering points are filtered out by corner response value thresholding, and the remaining candidate strong scattering points are target strong scattering points, which are used as seed points for line segment extraction of linear features composed of discrete strong scattering points in the image; all pixel points in the image are screened for amplitude maximum value points and points with an amplitude threshold greater than a set threshold to obtain a candidate growth point set; In the second aspect, for each seed point and candidate growth point, the principal direction of the seed point is obtained by weighted principal component analysis (PCA) based on the pixel amplitude values of the pixel points in the constructed annular neighborhood window, and the principal direction of the seed point is used to provide directional constraint for the extraction of linear features composed of discrete strong scattering points in the image; In the third aspect, each seed point is taken as a starting point, and the candidate growth points are taken as growth objects, and the line segments are grown in the principal direction and the opposite direction of the seed point; new growth points are found from the candidate growth points based on spatial distance, directional consistency and fitting direction vector change constraint, and the line segment growth is performed, the line segment direction is updated by straight line fitting of the current growth point set, until each line segment growth terminates, a candidate line segment set is obtained, and thus the complete line segment extraction of the linear features composed of discrete strong scattering points in the image is realized.

[0013] In the fourth aspect, the directional consistency of each candidate line segment in the candidate line segment set is verified, and the candidate line segments with directional consistency are screened out to obtain a target line segment set, and thus the high-precision line segment extraction of the linear features composed of discrete strong scattering points in the image is realized. In summary, the method provided in the application can effectively suppress the influence of noise in the image and improve the precision of line segment extraction in the process of extracting line segments of linear features composed of discrete strong scattering points in the image, so as to realize high-precision extraction of linear structures of buildings. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments description. Obviously, the drawings in the following description only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0015] In order to more completely understand the application and its beneficial effects, the following will be described in conjunction with the drawings, wherein the same reference numerals in the following description represent the same parts.

[0016] Figure 1 The flowchart of a high-resolution SAR satellite image line segment extraction method provided in the embodiments of the application.

[0017] Figure 2 The high-resolution SAR satellite image legend of a certain city area provided in the embodiments of the application.

[0018] Figure 3 The seed point diagram obtained based on the image shown in the embodiments of the application. Figure 2

[0019] Figure 4 The annular neighborhood window diagram provided in the embodiments of the application.

[0020] Figure 5 The estimation diagram of the main direction of the seed point obtained based on the image shown in the embodiments of the application. Figure 2

[0021] The grouping growth result diagram obtained based on the image shown in the embodiments of the application. Figure 6 Figure 2

[0022] Figure 7 The effective line segment extraction result diagram obtained based on the image shown in the embodiments of the application. Figure 2

[0023] The line segment extraction comparison diagram of the image shown in the embodiments of the application by using different methods. Figure 8 Figure 2

[0024] Figure 9 ​​​​​A schematic diagram of a high-resolution SAR satellite image line segment extraction system provided in an embodiment of the present application.

[0025] Figure 10 A structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0027] Figure 1 A flowchart of a high-resolution SAR satellite image line segment extraction method provided in an embodiment of the present application. Please refer to Figure 1 The embodiments of the present application provide a high-resolution SAR satellite image line segment extraction method, which includes steps 100 to 300, which will be described in detail below.

[0028] Step 100, acquiring a high-resolution SAR satellite image, processing the image into an amplitude image; based on constant false alarm rate detection, non-maximum suppression and diagonal point response thresholding, obtaining a seed point set; based on non-maximum suppression and adaptive amplitude thresholding of the amplitude image, obtaining a candidate growth point set.

[0029] Step 200, taking each point in the seed point set and the candidate growth point set as the center, constructing a ring neighborhood window, based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point, obtaining the main direction of each seed point in the seed point set and the main direction of each candidate growth point in the candidate growth point set through weighted principal component analysis.

[0030] Step 300, for each seed point, based on the main direction of the seed point and the main direction of the candidate growth point, using spatial distance and direction consistency constraints to determine a first growth point from the candidate growth point set, generating an initial fitting line segment; step length growth step: performing step length growth in the positive and negative directions of the initial fitting line segment or the current fitting line segment, obtaining a current step length growth line segment; second growth point determination step: based on the end points of the current step length growth line segment, using spatial distance, direction consistency and fitting direction vector change constraints to determine a current second growth point from the candidate growth point set, generating a current fitting line segment; repeating the step length growth step and the current second growth point determination step until no new second growth point can be found.

[0031] The method provided in the embodiments of the present application, in the first aspect, uses constant false alarm rate (CFAR) detection to find candidate strong scattering points from the image; candidate growth points are screened out from the candidate strong scattering points through non-maximum suppression (NMS); further, noise points in the candidate strong scattering points are filtered out through corner response value thresholding, and the remaining candidate strong scattering points are target strong scattering points, which are used as seed points for line segment extraction of linear features composed of discrete strong scattering points in the image; all pixel points in the image are screened to obtain a candidate growth point set, which includes amplitude maximum value points and points with an amplitude threshold greater than a set threshold. In the second aspect, for each seed point and candidate growth point, the main direction of the seed point is obtained through weighted principal component analysis (PCA) based on the pixel amplitude values of the pixel points in the constructed annular neighborhood window, and the main direction of the seed point is used to provide directional constraint for the extraction of linear features composed of discrete strong scattering points in the image. In the third aspect, each seed point is used as a starting point, and the candidate growth points are used as growth objects, and the line segments are grown in the main direction and the opposite direction of the seed point; new growth points are found from the candidate growth points based on spatial distance, direction consistency and fitting direction vector change constraint, the line segment direction is updated through straight line fitting of the current growth point set, and the process is repeated until each line segment growth terminates, thereby obtaining a candidate line segment set, and the complete line segment extraction of the linear features composed of discrete strong scattering points in the image is realized.

[0032] In the fourth aspect, the direction consistency of each candidate line segment in the candidate line segment set is verified, and the candidate line segments with direction consistency are screened out to obtain a target line segment set, thereby realizing high-precision line segment extraction of the linear features composed of discrete strong scattering points in the image. In summary, the method provided in the embodiments of the present application can effectively suppress the influence of noise in the image and improve the precision of line segment extraction in the process of line segment extraction of linear features composed of discrete strong scattering points in the image, thereby realizing high-precision extraction of building linear structures.

[0033] It should be noted that in the embodiments of the present application, the line segment precision reflects the approximation degree of the line segment extraction result to the true structure.

[0034] The following describes Figure 1 When performing high-resolution SAR satellite image line segment extraction processing in the embodiments, further optional specific embodiments of each step are described.

[0035] In step 100, a high-resolution SAR satellite image is obtained, and the image is processed into an amplitude image; a seed point set is obtained based on constant false alarm rate detection, non-maximum suppression and corner response value thresholding; and a candidate growth point set is obtained based on non-maximum suppression and an adaptive amplitude threshold of the amplitude image.

[0036] In some embodiments, the step of screening a set of seed points from the amplitude image based on constant false alarm rate detection, non-maximum suppression and diagonal point response thresholding, comprises: Detecting each pixel point in the amplitude image by a detection window; for each pixel point to be detected, generating an ordered sequence of all pixel points in the non-protected area of the detection window according to pixel amplitude values, obtaining a local pixel amplitude threshold value according to the ordered sequence and an empirical quantile ratio, and determining the pixel point to be detected as a candidate target pixel point if the pixel amplitude value of the pixel point to be detected is greater than the local pixel amplitude threshold value, thereby obtaining a set of candidate target pixel points; Comparing the pixel amplitude values of all candidate target pixel points in a third set neighborhood of each candidate target pixel point, and only retaining the candidate target pixel point with the largest filtered pixel amplitude value as a candidate feature point, thereby obtaining a set of candidate feature points; Calculating the corner point response value of each candidate feature point; comparing the corner point response value of each candidate feature point with a corner point response threshold value, and retaining the candidate feature point with a corner point response value greater than the set threshold value as a seed point, thereby obtaining a set of seed points.

[0037] It can be understood that in the present embodiment, candidate target pixel points can be obtained by constant false alarm rate detection, which are strong scattering points in the image; a set of candidate feature points can be obtained by non-maximum suppression, which are used to highlight isolated scattering structures; and noise points can be filtered out by non-maximum suppression, which are used for preliminary positioning of irregular structures in the image.

[0038] It can be understood that the detection window is a local neighborhood closely surrounding the pixel point to be detected, which includes a protected area and a non-protected area, the protected area being closely arranged around the pixel point to be detected, and the non-protected area being closely arranged around the protected area.

[0039] For example, please refer to Figure 2 , Figure 2 For a high-resolution SAR satellite image of a certain urban area. In constant false alarm rate detection, each pixel point in the image is detected, and a detection window is established around the pixel point to be detected; in the detection window, the pixel points in the protected area are removed, the pixel points in the non-protected area are retained, the pixel points in the non-protected area are arranged in ascending order according to the pixel amplitude values, and the local pixel amplitude threshold value of the pixel point to be detected is obtained in combination with the empirical quantile ratio. The calculation formula of the local pixel amplitude threshold value is: ， wherein: is the pixel point to be detected a local pixel magnitude threshold value of the pixel point to be detected; a set of pixel magnitude values of the pixel points within the non-protection region; a set of pixel points within the detection window of the pixel point to be detected a set of pixel points within the detection window of the pixel point to be detected a set of pixel points within the non-protection region in the detection window of the pixel point to be detected a set of pixel points within the non-protection region in the detection window of the pixel point to be detected a set of pixel points within the non-protection region in the detection window of the pixel point to be detected a set of pixel points within the non-protection region in the detection window of the pixel point to be detected a set of pixel points within the non-protection region in the detection window of the pixel point to be detected a sorting function for obtaining a sorted sequence; an index operator for extracting an element at a specific position from the sorted sequence, a floor function, an empirical quantile ratio, a number of pixel points within the non-protection region.

[0040] If the pixel magnitude value of the pixel point to be detected satisfies , the pixel point to be detected is determined as a constant false alarm detection response point, that is, a candidate target pixel point, that is, a strong scattering point.

[0041] For example, the empirical quantile ratio can be 90%. The third set of neighborhood of the candidate target pixel point can be a 3x3 neighborhood.

[0042] In some embodiments, the step of thresholding the corner point response value includes: calculating the corner point response value of each candidate feature point; comparing the corner point response value of each candidate feature point with a corner point response threshold value, and retaining the candidate feature points with a corner point response value greater than the set threshold value as seed points.

[0043] It can be understood that in this embodiment, the quality of the candidate feature points is evaluated by calculating the corner point response value and comparing it with the set threshold value, and the candidate feature points with low corner point response value due to noise and clutter are removed, and the stable and reliable strong scattering candidate feature points are retained as seed points for line segment growth.

[0044] For example, the corner point response value can be obtained by any one of the Harris response function, Shi-Tomasi corner detection, and FAST corner detection. In the specific embodiments of the present application, Figures 2 to 7 the corner point response value is obtained by the Harris response function.

[0045] For example, when the corner point response value of the candidate feature point is calculated by the Harris response function, the expression of the Harris response function is as follows: , , wherein, R is the corner response value of the candidate feature point to be calculated, M is a structure tensor calculated from the local image gradient of the candidate feature point to be calculated, is an empirical coefficient, . and respectively represent the gradients in the x direction and the y direction. Please refer to Figure 3 , Figure 3 is a seed point diagram provided in the embodiment of the application based on the image shown in FIG. 2. Figure 2

[0046] In some embodiments, based on the non-maximum suppression and the adaptive amplitude threshold of the amplitude image, the step of obtaining the candidate growth point set comprises: comparing the pixel amplitude values in the fourth set neighborhood of each pixel point of the amplitude image, and only retaining the pixel point with the maximum pixel amplitude value to obtain a set of candidate pixel points; further screening the candidate growth point set by using the adaptive amplitude threshold of the overall amplitude image to obtain an optimized candidate growth point set; The calculation formula of the adaptive amplitude threshold is: , wherein, is the adaptive amplitude threshold, is a control factor, is the background mean value of the overall image, is the background standard deviation of the overall image. Exemplarily, The value range of is greater than or equal to 2.0 and less than or equal to 3.0.

[0047] Exemplarily, for all the response points obtained by the non-maximum suppression , the response points with the pixel amplitude value satisfying are retained as the candidate growth points.

[0048] Exemplarily, the fourth set neighborhood of each pixel point can be a region composed of the four directly adjacent pixel points above, below, left and right of the each pixel point.

[0049] It can be understood that in the embodiment, the seed points for line segment growth are further screened from the candidate feature point set obtained by the non-maximum suppression. This step aims to eliminate false response points with low signal-to-noise ratio by calculating the adaptive amplitude threshold of the background noise of the overall image, so as to improve the accuracy and reliability of subsequent growth.

[0050] ​​It can be understood that the background mean value of the whole image and the background standard deviation of the whole image may be obtained by: determining the most stable background region in the whole image by searching for the window with the minimum local standard deviation in the whole image; obtaining the background mean value of the whole image and the background standard deviation of the whole image from the determined most stable background region.

[0051] In step 200, a ring neighborhood window is constructed centered on each point in the seed point set and the candidate growth point set, and the principal direction of each seed point in the seed point set and the principal direction of each candidate growth point in the candidate growth point set are obtained by weighted principal component analysis based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point.

[0052] It can be understood that each seed point has a principal direction. The principal direction of the seed point is used to indicate the principal direction of the neighborhood gray scale structure of the seed point, to provide a stable and locally dominant direction angle value for the seed point, to provide directional guidance for the subsequent line segment growth of the seed point, and to provide directional constraints for the subsequent growth of the seed point.

[0053] It can be understood that each candidate growth point has a principal direction, which is used to indicate the principal direction of the neighborhood gray scale structure of the candidate growth point, to provide constraints for the direction consistency determination of the candidate point in the line segment growth process, and to serve as a basis for estimating and verifying the direction of the subsequent fitting line segment.

[0054] In some embodiments, the step of obtaining the principal direction of each seed point in the seed point set and the principal direction of each candidate growth point in the candidate growth point set based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point by weighted principal component analysis includes: a weighted average coordinate calculation step: obtaining the relative coordinates and corresponding pixel amplitude values of a plurality of sample pixel points in the ring neighborhood window for the current seed point or the current candidate growth point, the relative coordinates being the coordinates of the plurality of sample pixel points relative to the current seed point; calculating the weighted average coordinates of the plurality of sample pixel points using the relative coordinates and corresponding pixel amplitude values; a weighted covariance matrix construction step: centering the relative coordinates of all the plurality of sample pixel points relative to the weighted average coordinates; calculating the weighted covariance matrix using the centered coordinates and corresponding pixel amplitude values; a matrix eigenvalue decomposition step: performing eigenvalue decomposition on the weighted covariance matrix to obtain a feature vector matrix composed of feature vectors and a feature value diagonal matrix composed of feature values; The main direction determining step: selecting the largest eigenvalue from the eigenvalue diagonal matrix, and defining the eigenvector corresponding to the eigenvalue in the eigenvector matrix as the target eigenvector; taking the target eigenvector as the main direction of the current seed point.

[0055] Exemplarily, a ring neighborhood window with a center of the seed point, an inner diameter of =7, and an outer diameter of is constructed. Please refer to Figure 4 , Figure 4 for a schematic diagram of the ring neighborhood window provided in the embodiments of the present application.

[0056] All pixel points in the ring neighborhood window satisfying are taken as sample pixel points, wherein is the coordinate of the target seed point, is the coordinate of the pixel point.

[0057] Exemplarily, the sample pixel point set can be defined as: , wherein is the sample pixel point set, is the coordinate of the target seed point, is the coordinate of the pixel point, is the inner diameter of the ring neighborhood window, is the outer diameter of the ring neighborhood window.

[0058] Exemplarily, the weighted mean coordinate of the sample pixel point set is calculated according to the following formula: , wherein is the weighted mean coordinate, is the relative coordinate of the i-th sample pixel point, , is the number of sample pixel points, is the pixel amplitude value of the i-th sample pixel point. The weighted covariance matrix is calculated according to the following formula: , wherein is the weighted covariance matrix, is the weighted mean coordinate, denotes the transpose of the matrix, is the relative coordinate of the i-th sample pixel point, , is the number of sample pixel points, is the pixel amplitude value of the i-th sample pixel point. ​​​a pixel amplitude value of a sample pixel point; The step of performing eigen-decomposition on the weighted covariance matrix comprises: , wherein, is the weighted covariance matrix, V is the eigenvector matrix, D is the eigenvalue diagonal matrix, represents the transpose of a matrix; The main direction of the target seed point is expressed as: , wherein, is the eigenvalue with the largest value selected from the eigenvalue diagonal matrix D and the corresponding eigenvector in the eigenvector matrix V ; is the component on the axis, is the component on the axis.

[0059] It can be understood that each sub-point has a main direction. The main direction of a seed point is used to indicate the main direction of the gray structure of the neighborhood of the seed point, to provide a stable and locally dominant direction angle value for the seed point, to provide a directional guidance basis for subsequent line segment growth of the seed point, and to provide a directional constraint for subsequent growth of the seed point. Please refer to Figure 5 , Figure 5 is the estimation map of the main direction of the seed point obtained based on the image shown in Figure 2 . Figure 5 The red arrow shown in

[0060] Step 300: for each sub-point, based on the main direction of the seed point and the main direction of the candidate growth point, a first growth point is determined from the candidate growth point set by using spatial distance and directional consistency constraints, and an initial fitting line segment is generated; step of step growth: step growth is performed in the positive and negative directions of the initial fitting line segment or the current fitting line segment to obtain a current step growth line segment; step of second growth point determination: based on the end point of the current step growth line segment, a current second growth point is determined from the candidate growth point set by using spatial distance, directional consistency and fitting direction vector change constraints, and a current fitting line segment is generated; the step of step growth and the step of current second growth point determination are repeated until no new second growth point can be found.

[0061] It can be understood that step 300 is a line segment growing step. Exemplarily, the specific steps of line segment growing can include: a current seed point determining step 310, an initial candidate growing point determining step 320, a step length growing step 330, a second growing point determining step 340, a current fitting growing step 350, and an iterative growing step 360.

[0062] The current seed point determining step 310: from the seed point set, the current seed point is determined according to the size of the corner point response value of the seed point.

[0063] The initial candidate growing point determining step 320: in the first set neighborhood of the current seed point, search for a candidate growing point satisfying the first condition as the initial candidate growing point, the first condition being that the included angle between the main direction of the candidate growing point and the main direction of the seed point is less than the first angle threshold, and the Euclidean distance between the candidate growing point and the target seed point is the smallest.

[0064] The step length growing step 330: based on the fitting line segment, which is any one of the initial fitting line segment and the current fitting line segment, grow in the positive and negative directions of the fitting line segment according to the step length , , is the step length coefficient, , the length of the initial fitting line segment, to obtain the current step length growing line segment.

[0065] The second growing point determining step 340: in the second set neighborhood of the end point of the current step length growing line segment, find a candidate growing point that has not been added to the current point set and satisfies the second condition as the second growing point, the second condition being that the vertical distance from the candidate growing point to the current fitting line segment is less than the vertical distance threshold, the vertical distance threshold being greater than 1 and less than 3; the included angle difference between the main direction of the candidate growing point and the direction of the current fitting line segment is less than the second angle threshold, the second angle threshold being greater than 20° and less than 30°; and after adding the candidate growing point to the point set, an updated point set is obtained, a new fitting line segment is obtained by performing straight line fitting on the updated point set; the angle difference between the direction of the new fitting line segment and the direction of the previous fitting line segment is less than the third angle threshold, the third angle threshold being 20°.

[0066] The current fitting growing step 350: if the second growing point is found, add the second growing point to the current point set to obtain an updated point set; perform straight line fitting on the updated point set by the least square method to obtain the current fitting line segment; if the second growing point is not found, terminate the growing.

[0067] Iterative growth step 360: After obtaining the fitted line segment, return to the step growth step; if no candidate growth point that meets the second condition is found after returning to the step growth step, return to the step growth step again. When the total number of times the step growth step is returned reaches the preset number of returns, and no candidate growth point that meets the second condition is found, stop returning and terminate the growth.

[0068] Understandably, the set of points formed during the growth of all line segments, along with the orientation information of each point within that set, will serve as a candidate set for subsequent orientation consistency verification of the fitted line segments. Please refer to [link / reference]. Figure 6 , Figure 6 The embodiments provided in this application are based on Figure 2 The image shown is a diagram of the grouped growth results obtained from the image.

[0069] For example, the coordinate calculation formula for the endpoints of the current step-size growth segment is: ,in As the endpoint, For the current set of grown points The first in One point, For the current set of growth points The coordinates of the points; Step size, The unit direction vector of the fitted line segment obtained by fitting a straight line to the current set of points. The calculation formula is: , in, The slope of the fitted line segment obtained by fitting a straight line to the current point set; by fitting a straight line to the current point set, a linear model of the fitted line segment is obtained. ,in The x-axis is... The vertical axis is , The slope The intercept is used to obtain the unit direction vector of the fitted line segment. .

[0070] For example, line fitting methods can include least squares method, RANSAC line fitting, or orthogonal fitting method. This application Figures 2 to 7 In a specific embodiment, the least squares method is used for line fitting.

[0071] In some embodiments, for each seed point, the step of determining a first growth point from the set of candidate growth points based on the principal direction of the seed point and the principal direction of the candidate growth points, using spatial distance and orientation consistency constraints, includes: Determine the current seed point from the set of seed points; Based on the candidate growth point set, a candidate growth point that satisfies the spatial distance and direction consistency constraint conditions simultaneously is found in the first set neighborhood of the current seed point, and is determined as the first growth point; the spatial distance and direction consistency constraint conditions include: the angle between the main direction of the candidate growth point and the main direction of the current seed point is less than a first angle threshold; and the Euclidean distance between the candidate growth point and the current seed point is the smallest.

[0072] It can be understood that the above spatial distance constraint condition and direction consistency constraint condition are used to ensure the consistency of the first growth point and the seed point in spatial position and direction, and to ensure the continuity and accuracy of the line segment growth.

[0073] For example, the first angle threshold is greater than or equal to and less than or equal to .

[0074] The first set neighborhood of the seed point can be a neighborhood with a radius of constructed with the seed point as the center, Further, in some embodiments, the step of determining the current seed point from the seed point set includes: sorting each seed point in the seed point set according to the corner point response value from large to small, and determining the current seed point according to the sorting sequence.

[0075] In some embodiments, based on the end point of the current step length growth line segment, the spatial distance constraint, the direction consistency constraint, and the fitting direction vector change constraint are used to determine the current second growth point from the candidate growth point set, including: Based on the candidate growth point set, a candidate growth point that satisfies the spatial distance, direction consistency, and fitting direction vector change constraint conditions simultaneously is found in the second set neighborhood of the current step length growth line segment, and is determined as the current second growth point; wherein the spatial distance, direction consistency, and fitting direction vector change constraint conditions include: the perpendicular distance from the candidate growth point to the current fitting line segment is less than a perpendicular distance threshold, and the value of the perpendicular distance threshold is greater than 1 and less than 3; the angle difference between the main direction of the candidate growth point and the direction of the current fitting line segment is less than a second angle threshold; and After adding the candidate growth point to the updated point set, a new fitting line segment is obtained by performing straight line fitting; the angle difference between the direction of the new fitting line segment and the direction of the previous fitting line segment is less than a third angle threshold.

[0076] Exemplarily, the vertical distance threshold can be greater than 1 and less than 3. The second angle threshold and the third angle threshold can be greater than 20° and less than 30°.

[0077] It can be understood that the above spatial distance constraint condition, the direction consistency constraint condition and the fitting direction vector change constraint condition are used to ensure the stable growth of the line segment in the direction consistency, avoid the structure breaking, and ensure the stability and accuracy of the line segment extraction.

[0078] Further, in some embodiments, after the step of searching for the candidate growth point satisfying the spatial distance, the direction consistency and the fitting direction vector change constraint condition, the method further comprises: If no candidate growth point satisfying the spatial distance, the direction consistency and the fitting direction vector change constraint condition is found in the second set neighborhood, jumping to the step of step length growth, and when the jumping number reaches a first number threshold, terminating the growth; If the candidate growth point satisfying the spatial distance, the direction consistency and the fitting direction vector change constraint condition is found in the second set neighborhood, constructing the current seed point, the first growth point and all the found second growth points as a current point set, performing a straight line fitting based on the current point set, and generating a current fitting line segment.

[0079] Exemplarily, the preset return number can be set to 3 times.

[0080] Further, in some embodiments, after the step of searching for the candidate growth point satisfying the spatial distance, the direction consistency and the fitting direction vector change constraint condition, the method further comprises: If the candidate growth point satisfying the spatial distance, the direction consistency and the fitting direction vector change constraint condition is found in the second set neighborhood, the candidate growth point satisfying the spatial distance, the direction consistency and the fitting direction vector change constraint condition is determined as a candidate point; and the candidate point is checked, comprising: obtaining a current point set corresponding to the current fitting line segment; judging whether the candidate point belongs to the current point set; if yes, the candidate point is not determined as the current second growth point, and a new candidate growth point is searched; and if no, the candidate point is determined as the current second growth point.

[0081] In some embodiments, the step of growing in the positive and negative directions of the initial fitting line segment or the current fitting line segment to obtain the current step length growth line segment comprises: When the fitting line segment is the initial fitting line segment, growing in the positive and negative directions of the initial fitting line segment according to the step length , , is a step length coefficient, a length of the initial fitting line segment; when the fitting line segment is the current fitting line segment, growing along the positive and negative directions of the current fitting line segment by a step length , , a step length coefficient, a length of the initial fitting line segment.

[0082] exemplarily, the value range of the step length coefficient can be .

[0083] exemplarily, the step length is a multiple of the length of the initial fitting line segment. The step length is calculated according to the following formula: wherein, a step length coefficient, , a length of the initial fitting line segment, , denotes an initial candidate growth point, denotes a target seed point, denotes a displacement vector from to , and also denotes the length of the initial fitting line segment.

[0084] In some embodiments, after repeating the step of growing by a step length and the step of determining a current second growth point until the step of failing to find a new second growth point, the method further comprises: performing a direction consistency verification on the current fitting line segment obtained based on each seed point, comprising: based on the current seed point and the corresponding current fitting line segment, constructing the current seed point, the first growth point and all the found second growth points into a current point set; performing a straight line fitting on the current point set to obtain a current fitting direction; calculating an angle difference between the principal direction of each point in the current point set and the current fitting direction, counting the number of points with an angle difference less than a fourth angle threshold, counting the total number of points in the current point set, and obtaining a first ratio by dividing the number by the total number; if the first ratio is greater than or equal to a ratio threshold, determining that the current fitting line segment is a valid line segment and retaining it; if the first ratio is less than the ratio threshold, determining that the current fitting line segment is an invalid line segment and eliminating it.

[0085] exemplarily, the set of all points in the current fitting line segment is referred to as a current point set; for the current point set , first perform a linear fitting on the spatial position of each point in the current point set by using the least square method to obtain a current fitting straight line , the current fitting direction of the current fitting straight line is calculated For near-vertical straight lines, define .

[0086] For each point in the current point set, the principal direction of the point is calculated and the angle difference between the principal direction of the point and the current fitting direction of the fitting straight line is calculated : , wherein, denotes taking the minimum value, and |·| denotes taking the absolute value. If the angle difference of the point is less than a fourth angle threshold , wherein the value range of is greater than and less than , the point is considered to be consistent with the overall structural direction of the current fitting straight line, and the point is determined to be a direction-consistent point. It can be understood that the current fitting direction is used to represent the overall structural direction of the current fitting straight line.

[0087] The number of all direction-consistent points in the current point set is counted , and a first proportion, that is, the direction-consistency ratio of the current point set, is calculated wherein, is the direction-consistency ratio of the current point set, is the number of all direction-consistent points in the current point set, is the number of all points in the current point set.

[0088] If the direction-consistency ratio of the current point set , wherein is a ratio threshold, and exemplarily, the value of is 0.8, it is determined that the current fitting line segment is a valid line segment and is retained; if , it is determined that the current fitting line segment is an invalid line segment and is removed.

[0089] Finally, all current fitting line segments that pass the direction-consistency verification constitute the valid line-shaped structural feature set extracted from the image.

[0090] Figure 7 is an effective line segment extraction result image obtained based on the image shown in Figure 2 .

[0091] Figure 8For the comparative effect of LSDSAR, POE, Hough, EDLine algorithm on the same image. Among them, the LSDSAR (Line Segment Detector for SAR images) algorithm refers to the line segment detection algorithm for SAR images; the POE (Pixel Orientation Estimation) algorithm refers to the pixel direction estimation algorithm; the Hough algorithm refers to the Hough transform algorithm; the EDLines (Edge Drawing Lines) algorithm refers to the line segment detection algorithm based on edge drawing.

[0092] By combining Figure 8 With Figure 7 And Figure 1 The original drawing comparison shows that the line segment extracted by the method provided in the application embodiment has higher line segment accuracy and precision, and can obtain more complete and clear building structures, verifying the reliability of the method provided in the application embodiment in extracting linear structures under high-resolution SAR satellite images. The line segment accuracy reflects the proportion of effective line segments in the line segment extraction result; the line segment precision reflects the approximation degree of the line segment extraction result to the true structure.

[0093] Please refer to Figure 9 The application embodiment also provides a high-resolution SAR satellite image line segment extraction system, comprising a growth point determination module 901, a main direction determination module 902 of the growth point, and a line segment growth module 903.

[0094] The growth point determination module 901 is configured to: acquire a high-resolution SAR satellite image, process the image into an amplitude image; based on constant false alarm rate detection and non-maximum suppression, screen a candidate growth point set from the amplitude image; and by thresholding the corner point response value, screen a seed point set from the candidate growth point set.

[0095] The main direction determination module 902 of the growth point is configured to: take each point in the seed point set and the candidate growth point set as the center to construct a ring neighborhood window, and based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point, obtain the main direction of each seed point in the seed point set and the main direction of each candidate growth point in the candidate growth point set by weighted principal component analysis.

[0096] The line segment growing module 903 is configured to: for each seed point, determine a first growing point from the candidate growing point set based on the main direction of the seed point and the main direction of the candidate growing point by using the spatial distance and direction consistency constraint, and generate an initial fitting line segment; a step growing step: performing step growing in the positive and negative directions of the initial fitting line segment or the current fitting line segment to obtain a current step growing line segment; a second growing point determining step: determining a current second growing point from the candidate growing point set based on the end point of the current step growing line segment by using the spatial distance, the direction consistency, and the fitting direction vector change constraint, and generating a current fitting line segment; repeating the step growing step and the current second growing point determining step until a new second growing point cannot be found.

[0097] The embodiment of the present application further provides another line segment extraction method for high-resolution SAR satellite images, comprising: obtaining a high-resolution SAR satellite image, processing the image into an amplitude image, obtaining a seed point set based on constant false alarm rate detection, non-maximum suppression, and diagonal point response value thresholding, and obtaining a candidate growing point set based on non-maximum suppression and an adaptive amplitude threshold of the amplitude image; constructing a ring neighborhood window with each point in the seed point set and the candidate growing point set as the center, obtaining the main direction of each seed point in the seed point set and the main direction of each candidate growing point in the candidate growing point set by weighted principal component analysis based on the pixel amplitude values of the pixel points in the ring neighborhood window of each point; for each seed point, determining a first growing point from the candidate growing point set based on the main direction of the seed point and the main direction of the candidate growing point by using the spatial distance and direction consistency constraint, and generating a current fitting line segment based on the first growing point by using the spatial distance, the direction consistency, and the fitting direction vector change constraint; a new second growing point determining step: determining a new second growing point from the candidate growing point set based on the current second growing point and the current fitting line segment by using the spatial distance, the direction consistency, and the fitting direction vector change constraint, and generating a new fitting line segment; repeating the new second growing point determining step until a new second growing point cannot be found.

[0098] The embodiment of the present application further provides another line segment extraction system for high-resolution SAR satellite images, comprising: a growing point determining module configured to: obtain a high-resolution SAR satellite image, process the image into an amplitude image, obtain a seed point set based on constant false alarm rate detection, non-maximum suppression, and diagonal point response value thresholding, and obtain a candidate growing point set based on non-maximum suppression and an adaptive amplitude threshold of the amplitude image; The main direction determination module for growth points is used to: construct an annular neighborhood window centered on each point in the seed point set and the candidate growth point set; and obtain the main direction of each seed point in the seed point set and the main direction of each candidate growth point in the candidate growth point set through weighted principal component analysis based on the pixel amplitude value of the pixels in the annular neighborhood window of each point. The line segment growth module is used for: for each seed point, determining a first growth point from the set of candidate growth points based on the main direction of the seed point and the main direction of the candidate growth points, using spatial distance and direction consistency constraints; determining a second growth point from the set of candidate growth points based on the first growth point, using spatial distance, direction consistency, and fitting direction vector change constraints, and generating the current fitted line segment; the new second growth point determination step: determining a new second growth point from the set of candidate growth points based on the current second growth point and the current fitted line segment, using spatial distance, direction consistency, and fitting direction vector change constraints, and generating a new fitted line segment; repeating the new second growth point determination step until no new second growth point can be found.

[0099] In practical applications, the aforementioned system can be a terminal device or a chip applied to a terminal device. In this application, the system can implement the functions of multiple units through software, hardware, or a combination of both, enabling the system to execute the high-resolution SAR satellite image line segment extraction method steps provided in any of the above embodiments. Furthermore, the technical effects of each technical solution in this system can be referenced to the technical effects of the corresponding technical solutions in the high-resolution SAR satellite image line segment extraction method, and will not be elaborated upon further in this application.

[0100] Figure 10 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application.

[0101] Based on the hardware implementation of each unit in the above system, embodiments of this application also provide an electronic device, such as... Figure 8 As shown, the electronic device 10 includes a memory 110 and a processor 120. The memory 110 stores a computer program, and the processor 120 executes the computer program to implement the steps of the high-resolution SAR satellite image line segment extraction method provided in any of the above embodiments.

[0102] Of course, in practical applications, such as Figure 8 As shown, the various components in the electronic device 10 are coupled together via a bus system 130. It is understood that the bus system 130 is used to enable communication between these components. In addition to a data bus, the bus system 130 also includes a power bus, a control bus, and a status signal bus. However, for clarity, all buses are labeled as bus system 130 in the figure.

[0103] In practical applications, the processor can be at least one of an application specific integrated circuit (ASIC), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, and a microprocessor. It can be understood that, for different devices, the electronic device used to implement the functions of the processor can also be other devices, and the embodiments of the present application do not make specific limitations.

[0104] The memory can be a volatile memory (such as a random access memory (RAM)), a non-volatile memory (such as a read-only memory (ROM), a flash memory, a hard disk (HDD) or a solid state disk (SSD)), or a combination of the above types of memories, and provides instructions and data to the processor.

[0105] The electronic device described in the embodiments of the present application can be a terminal device, or a chip applied to a terminal device. The terminal device described in the embodiments of the present application can include terminal devices such as a mobile phone, a tablet computer, a notebook computer, a palm computer, and the like.

[0106] In the example embodiments, the embodiments of the present application also provide a computer readable storage medium, for example, a memory including a computer program, which can be executed by a processor of an electronic device to complete the steps of the foregoing method.

[0107] The embodiments of the present application also provide a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method described in any one of the embodiments of the present application.

[0108] Optionally, the computer program product can be applied to the electronic device in the embodiments of the present application, and the computer program instructions make the computer execute the corresponding processes realized by the electronic device in the various methods of the embodiments of the present application. For brevity, details are not repeated here.

[0109] The embodiments of the present application also provide a computer program.

[0110] Optionally, the computer program can be applied to the electronic device in the embodiments of the present application, and when the computer program runs on the computer, the computer is caused to execute the corresponding processes realized by the electronic device in each method of the embodiments of the present application. For brevity, details are not repeated here.

[0111] It should be understood that in the embodiments of the present application, the data related to user information and the like need to obtain the permission or consent of the user when the embodiments of the present application are applied to specific products or technologies, and the collection, use and processing of the related data need to comply with the relevant laws, regulations and standards of the country and region.

[0112] It should be understood that the terms used in the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms "a", "said" and "the" used in the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein means and includes any or all possible combinations of one or more associated listed items. In the present application, the expressions "have", "may have", "include" and "contain", or "may include" and "may contain" can be used herein to indicate the presence of the corresponding features (for example, elements such as numerical values, functions, operations or components), but do not exclude the presence of additional features.

[0113] It should be understood that although the terms first, second, third, etc. can be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other, and do not necessarily be used to describe a particular order or sequence. For example, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information, without departing from the scope of the present application.

[0114] The technical solutions described in the embodiments of the present application can be combined arbitrarily without conflict.

[0115] In several embodiments provided by the present application, it should be understood that the disclosed methods, devices and equipment can be implemented by other ways. The embodiments described above are only illustrative, for example, the division of units is only a logical function division, and actual implementation can have another division manner, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed components can be through some interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0116] The units described as separate parts above can or can not be physically separate, the parts shown as units can or can not be physical units, that is, can be located in one place or distributed on multiple network units; part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0117] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be realized in the form of hardware or in the form of hardware plus software functional unit.

[0118] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A high-resolution SAR satellite image line segment extraction method, characterized in that, The method comprises the following steps: acquiring a high-resolution SAR satellite image and processing the image into an amplitude image; obtaining a seed point set based on constant false alarm rate detection, non-maximum suppression and diagonal point response thresholding; obtaining a candidate growth point set based on non-maximum suppression and an adaptive amplitude threshold of the amplitude image; constructing a ring neighborhood window with each point in the seed point set and the candidate growth point set as the center, obtaining a main direction of each seed point in the seed point set and a main direction of each candidate growth point in the candidate growth point set through weighted principal component analysis based on pixel amplitude values of pixel points in the ring neighborhood window of each point; for each seed point, determining a first growth point from the candidate growth point set based on the main direction of the seed point and the main direction of the candidate growth point, generating an initial fitting line segment by using spatial distance and directional consistency constraints; a step growth step: performing step growth in the positive and negative directions of the initial fitting line segment or a current fitting line segment to obtain a current step growth line segment; a second growth point determination step: determining a current second growth point from the candidate growth point set based on end points of the current step growth line segment by using spatial distance, directional consistency and fitting direction vector change constraints to generate a current fitting line segment; repeating the step growth step and the current second growth point determination step until no new second growth point can be found.

2. The high-resolution SAR satellite image line segment extraction method according to claim 1, wherein the step of obtaining a seed point set based on constant false alarm rate detection, non-maximum suppression and diagonal point response thresholding comprises the following steps: detecting pixel points in the amplitude image one by one through a detection window; for each to-be-detected pixel point, generating a sorting sequence of all pixel points located in a non-protection area of the detection window according to pixel amplitude values, obtaining a local pixel amplitude threshold according to the sorting sequence and an empirical quantile ratio, and determining the to-be-detected pixel point as a candidate target pixel point if the pixel amplitude value of the to-be-detected pixel point is greater than the local pixel amplitude threshold to obtain a candidate target pixel point set; comparing pixel amplitude values of all candidate target pixel points in a third set neighborhood of each candidate target pixel point, and only retaining a candidate target pixel point with the largest pixel amplitude value as a candidate feature point to obtain a candidate feature point set; calculating an angle point response value of each candidate feature point, comparing the angle point response value of each candidate feature point with an angle point response threshold, and retaining a candidate feature point with an angle point response value greater than the set threshold as a seed point to obtain a seed point set.

3. The high-resolution SAR satellite image line segment extraction method according to claim 1, wherein the step of obtaining a candidate growth point set based on non-maximum suppression and an adaptive amplitude threshold of the amplitude image comprises the following steps: comparing pixel amplitude values in a fourth set neighborhood of each pixel point in the amplitude image, and only retaining a pixel point with the largest pixel amplitude value to obtain a to-be-selected pixel point set. ​ ​ Further filter the candidate pixel point set by using an adaptive amplitude threshold of the whole amplitude image to obtain an optimized candidate growing point set; The calculation formula of the adaptive amplitude threshold is: ， wherein, is an adaptive amplitude threshold, is a control factor, is a background mean value of the whole image, is a background standard deviation of the whole image.

4. The high-resolution SAR satellite image line segment extraction method of claim 1, characterized in that, The steps of obtaining the main direction of each seed point in the seed point set and the main direction of each candidate growing point in the candidate growing point set by using weighted principal component analysis based on the pixel amplitude values of the pixel points in the annular neighborhood window of each point include: A weighted average value coordinate calculation step: obtaining the relative coordinates and corresponding pixel amplitude values of a plurality of sample pixel points in the annular neighborhood window relative to the current seed point or the current candidate growing point, the relative coordinates being the coordinates of the plurality of sample pixel points relative to the current seed point; calculating the weighted average value coordinates of the plurality of sample pixel points by using the relative coordinates and corresponding pixel amplitude values; A weighted covariance matrix construction step: centering the relative coordinates of all the plurality of sample pixel points relative to the weighted average value coordinates; calculating the weighted covariance matrix by using the centered coordinates and corresponding pixel amplitude values; A matrix eigenvalue decomposition step: performing eigenvalue decomposition on the weighted covariance matrix to obtain a feature vector matrix composed of feature vectors and a feature value diagonal matrix composed of feature values; A main direction determination step: selecting the feature value with the largest value from the feature value diagonal matrix, and defining the feature vector corresponding to the feature value in the feature vector matrix as a target feature vector; taking the target feature vector as the main direction of the current seed point.

5. The high-resolution SAR satellite image line segment extraction method of claim 1, characterized in that, The steps of determining the first growing point from the candidate growing point set based on the main direction of the seed point and the main direction of the candidate growing point by using spatial distance and direction consistency constraints for each seed point include: Determining the current seed point from the seed point set; Based on the candidate growing point set, finding the candidate growing point that meets the spatial distance and direction consistency constraint conditions in the first set neighborhood of the current seed point to determine the first growing point; the spatial distance and direction consistency constraint conditions include: The angle between the main direction of the candidate growing point and the main direction of the current seed point is less than a first angle threshold; and The Euclidean distance between the candidate growing point and the current seed point is the smallest.

6. The high-resolution SAR satellite image line segment extraction method of claim 1, characterized in that, The steps of determining the current second growing point from the candidate growing point set based on the end point of the current step length growing line segment by using spatial distance constraints, direction consistency constraints and fitting direction vector change constraints include: Based on the candidate growth point set, a candidate growth point that simultaneously satisfies the spatial distance, the direction consistency and the fitting direction vector change constraint condition is searched in a second set neighborhood of the current step growth line segment, and is determined as a current second growth point; wherein the spatial distance, the direction consistency and the fitting direction vector change constraint condition include: A perpendicular distance of the candidate growth point to the current fitting line segment is less than a perpendicular distance threshold value; An included angle difference between a main direction of the candidate growth point and a direction of the current fitting line segment is less than a second angle threshold value; and An updated point set obtained after the candidate growth point is added to the point set is subjected to straight line fitting, and a new fitting growth line segment is obtained; an angle difference between a direction of the new fitting growth line segment and a direction of the previous fitting growth line segment is less than a third angle threshold value.

7. The high-resolution SAR satellite image line segment extraction method according to claim 6, further comprising: after the step of searching for the candidate growth point that simultaneously satisfies the spatial distance, the direction consistency and the fitting direction vector change constraint condition, if no candidate growth point that simultaneously satisfies the spatial distance, the direction consistency and the fitting direction vector change constraint condition is searched in the second set neighborhood, jumping to the step of step growth, and when a jumping number reaches a first number threshold value, terminating growth; if the candidate growth point that simultaneously satisfies the spatial distance, the direction consistency and the fitting direction vector change constraint condition is searched in the second set neighborhood, constructing the current seed point, the first growth point and all the searched second growth points into a current point set, and based on the current point set, performing straight line fitting to generate a current fitting line segment.

8. The high-resolution SAR satellite image line segment extraction method according to claim 1, wherein the step of performing step growth in the positive and negative directions of the initial fitting line segment or the current fitting line segment to obtain the current step growth line segment comprises:

9. The high-resolution SAR satellite image line segment extraction method according to claim 1, further comprising, after the step of repeating the step of step growth and the step of determining the current second growth point until no new second growth point is searched: performing direction consistency verification on the current fitting line segment obtained based on each kind of sub-point, including: When the fitting line segment is an initial fitting line segment, in both positive and negative directions of the initial fitting line segment, the step grows, , is a step coefficient, is a length of the initial fitting line segment; When the fitting line segment is a current fitting line segment, in both positive and negative directions of the current fitting line segment, a step grows, , is a step coefficient, is a length of an initial fitting line segment. constructing the current seed point, the first growth point and all the searched second growth points into a current point set based on the current seed point and the corresponding current fitting line segment; performing straight line fitting on the current point set to obtain a current fitting direction; calculating an angle difference between a main direction of each point in the current point set and the current fitting direction, counting a number of points with an angle difference less than a fourth angle threshold value, counting a total number of points in the current point set, and obtaining a first ratio by dividing the number by the total number; if the first ratio is greater than or equal to a ratio threshold value, determining that the current fitting line segment is a valid line segment and retaining the current fitting line segment; if the first ratio is less than the ratio threshold value, determining that the current fitting line segment is an invalid line segment and eliminating the current fitting line segment. ​ ​ 10. A high-resolution SAR satellite image line segment extraction method, characterized in that, ​ Acquire a high-resolution SAR satellite image, and process the image into an amplitude image; obtain a seed point set based on constant false alarm rate detection, non-maximum suppression, and diagonal point response thresholding; obtain a candidate growth point set based on non-maximum suppression and adaptive amplitude thresholding of the amplitude image; construct a ring neighborhood window centered on each point in the seed point set and the candidate growth point set, and obtain a main direction of each seed point in the seed point set and a main direction of each candidate growth point in the candidate growth point set by weighted principal component analysis based on pixel amplitude values of pixel points in the ring neighborhood window of each point; for each seed point, determine a first growth point from the candidate growth point set based on the main direction of the seed point and the main direction of the candidate growth point by using spatial distance and directional consistency constraints; determine a second growth point from the candidate growth point set based on the first growth point by using spatial distance, directional consistency, and fitting direction vector change constraints, and generate a current fitting line segment; determine a new second growth point from the candidate growth point set based on the current second growth point and the current fitting line segment by using spatial distance, directional consistency, and fitting direction vector change constraints, and generate a new fitting line segment; repeat the new second growth point determination step until a new second growth point cannot be found.

11. An electronic device, comprising: A device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the high-resolution SAR satellite image line segment extraction method provided in any one of claims 1 to 10 when executing the computer program.

Images