Target Detection Method, Device, Equipment and Storage Medium Based on High-Resolution SAR

Abstract

The present application discloses a target detection method, device, equipment and storage medium based on high-resolution SAR, relating to the technical field of target detection and classification. The method includes: obtaining a sliced image based on high-resolution SAR that contains sea surface ship targets, and determining its brightness mean value, as well as the total row brightness value and the total column brightness value; screening out the target row with the maximum brightness in the total row brightness value and the target column with the maximum brightness in the total column brightness value to determine the boundary positions of the corresponding scattering sidelobe suppression regions; based on preset conditions, using the interval between the boundary positions to determine the target sliced image that needs to be processed for scattering sidelobe suppression, and using a target detection algorithm to process it to obtain a target binary image; finally, determining the minimum circumscribed rectangle of the ship target in the target binary image. Through the technical solution of the present application, the trailing phenomenon caused by scattering sidelobes in the high-resolution SAR sea surface ship target image can be solved to realize the extraction of the minimum circumscribed rectangle of the target.

Description

Technical Field

[0001] The present invention relates to the technical field of target detection and classification, and particularly relates to a target detection method, device, equipment and storage medium based on high-resolution SAR. Background Art

[0002] The ocean is an important protection scope of national rights and interests, accounting for a large proportion of the national territory area. With the development of the ocean informatization construction of various countries, in the aspect of sea area monitoring, sea vessels are the key monitoring objects, which are of great significance for maritime traffic, fishery monitoring and national defense security. SAR (Synthetic Aperture Radar) has the characteristics of all-weather, all-day and long-distance microwave imaging, and has unique advantages in the monitoring of sea vessels, especially in the monitoring of emergencies.

[0003] The main task of sea vessel monitoring based on high-resolution SAR images is to detect vessel targets from SAR images and even achieve vessel classification based on target features. In this process, the range and scale information of vessels are very important for the determination of target properties. Generally speaking, for the extraction and classification of vessel target information based on SAR images, the primary problem to be solved is the definition of the target area range, which is transformed into a mathematical problem, that is, how to accurately extract the minimum bounding rectangle (MBR) of the target area. At present, in the research field of SAR image vessel detection and classification, the research on this problem is not prominent. The reasons are roughly as follows: on the one hand, due to the restriction of the low resolution of SAR images in previous research, the focus is often only on vessel detection, that is, as long as the target is detected, little attention is paid to the target scale or only rough extraction of the target scale information is needed; on the other hand, with the rapid development of high-resolution SAR systems, the research on sea vessel classification based on high-resolution SAR images is still in its infancy, and currently people pay more attention to the research on target features and classification algorithms. However, accurate and rapid extraction of the MBR of vessel targets is very important for vessel classification applications, which will seriously affect the accuracy of subsequent target feature extraction and thus affect the classification results.

[0004] At present, for the method of extracting the target MBR based on SAR image segmentation or target detection, the general processing flow is as follows: First, perform threshold-based target segmentation or target detection on the SAR target slice image to obtain a binary image of the target and the background; then, in the obtained binary images of the target and the background, use mathematical methods to calculate the MBR of the target area. Generally speaking, this extraction method usually has little impact on low-resolution SAR images. However, when applied to high-resolution SAR images, due to the prominent problem of SAR target scattering sidelobes, that is: in high-resolution SAR images, strong scattering parts of the target often form strong "one-word" or "cross" trailing phenomena, as Figure 1 shown, where (a) is the "cross" trailing phenomenon of the scattering sidelobe, and (b) is the "one-word" trailing phenomenon of the scattering sidelobe. As a result, in the segmented binary target image, the target area is often interfered by the "one-word" or "cross" scattering sidelobe trailing. At this time, directly extracting the MBR will seriously deviate from the MBR of the real target. To overcome this problem, there are usually two solutions. One is to preprocess the SAR ship target slice image, that is, reduce the influence of the scattering sidelobe through scattering sidelobe suppression; the other is to use morphological processing methods to process the target segmentation binary image to eliminate the "one-word" or "cross" scattering trailing in the binary image, and on this basis, extract the MBR of the ship target. Among them, due to the problems of processing difficulty and complexity, the former method has relatively little research. Individual literatures either use image enhancement algorithms or iterative search processing methods; while most people use the latter processing method, that is, remove the scattering sidelobe interference in the target segmentation binary image through morphological dilation, erosion and other algorithms.

[0005] In the prior art, there are the following technical defects when extracting the MBR of SAR sea ships: (1) Using image enhancement algorithms to preprocess SAR ship target images will cause the synchronous enhancement or weakening of the target area and sidelobe scattering interference, and it is difficult to achieve the effect of eliminating sidelobe scattering; (2) Using the iterative search method to process sidelobe scattering has a large calculation time, and in the later stage of iterative processing, as the sidelobe scattering interference weakens, other strong scattering parts in the ship target area will be mis-searched and processed; (3) When using the morphological processing method for target segmentation binary images to eliminate sidelobe scattering interference, it will inevitably expand or erode the real target area at the same time, resulting in serious distortion of the target MBR.

[0006] In summary, how to quickly and accurately extract high-resolution SAR sea ship targets that meet the requirements of rapid application, and the minimum circumscribed rectangle of the target in the case of strong scattering sidelobes is a problem to be solved at present. Summary of the Invention

[0007] In view of this, the object of the present invention is to provide a target detection method, device, equipment and storage medium based on high-resolution SAR, which can quickly and accurately extract high-resolution SAR sea vessel targets that meet the requirements of rapid application, and the minimum circumscribed rectangle of the target under the condition of strong scattering sidelobes. The specific scheme is as follows:

[0008] In a first aspect, the present application discloses a target detection method based on high-resolution SAR, including:

[0009] Obtain a sliced image based on high-resolution SAR containing sea vessel targets, and determine the brightness mean value of the sliced image, the total row brightness value of each row in the sliced image, and the total column brightness value of each column in the sliced image;

[0010] Select the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values, and determine the boundary positions of the corresponding scattering sidelobe suppression regions respectively based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column;

[0011] Based on preset conditions, use the interval between the boundary positions to determine the target sliced image that needs to be subjected to scattering sidelobe suppression processing on the sliced image, and process the processed target sliced image using a target detection algorithm to obtain a target binary image;

[0012] Determine the minimum circumscribed rectangle of the sea vessel target according to the region where the sea vessel target is located in the target binary image, so as to detect the sea vessel target.

[0013] Optionally, determining the brightness mean value of the sliced image includes:

[0014] Obtain the width information and height information of the sliced image;

[0015] Use the width information and the height information to determine the total number of pixels of the sliced image;

[0016] Based on the total number of pixels and the gray levels of all pixels in the sliced image, determine the brightness mean value of the sliced image.

[0017] Optionally, determining the total row brightness value of each row and the total column brightness value of each column in the sliced image includes:

[0018] Perform row-by-row statistics on the sliced image, and determine the total row brightness value of each row in the sliced image based on the height information and the gray levels of the pixels in each row of the sliced image;

[0019] Perform column-by-column statistics on the sliced image, and determine the total column brightness value of each column in the sliced image based on the width information and the gray level of each column of pixels in the sliced image.

[0020] Optionally, the determining the boundary positions of the respective scattering sidelobe suppression regions based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column includes:

[0021] Construct a threshold relationship for defining the range of the scattering sidelobe suppression region;

[0022] Based on the threshold relationship, use the brightness mean value and the brightness value of the target row to determine a first result threshold, and use the brightness mean value and the brightness value of the target column to determine a second result threshold;

[0023] Centered on the target row, compare the total row brightness values of other rows within a preset pixel range on both sides of the target row with the first result threshold. When the total row brightness value of the other rows is less than the first result threshold, determine the position of the other rows as the boundary position of the scattering sidelobe suppression region;

[0024] Centered on the target column, compare the total column brightness values of other columns within the preset pixel range on both sides of the target column with the second result threshold. When the total column brightness value of the other columns is less than the second result threshold, determine the position of the other columns as the boundary position of the scattering sidelobe suppression region.

[0025] Optionally, after determining the target sliced image that needs to be subjected to scattering sidelobe suppression processing on the sliced image based on the preset condition using the interval between the boundary positions, it further includes:

[0026] When the interval between the boundary positions is less than a preset pixel value, it is determined that there are scattering sidelobes between the boundary positions, and then the sliced image is subjected to scattering sidelobe suppression processing to obtain a target sliced image.

[0027] Optionally, after determining the target sliced image that needs to be subjected to scattering sidelobe suppression processing on the sliced image based on the preset condition using the interval between the boundary positions, it further includes:

[0028] Based on the brightness value of the target row and the boundary position corresponding to the target row, determine the pixel values for replacing the image rows within the scattering sidelobe suppression region to perform the scattering sidelobe suppression processing on the image rows;

[0029] Determine the pixel values for replacing the image columns within the scattered sidelobe suppression region based on the brightness values of the target columns and the boundary positions corresponding to the target columns, so as to perform the scattered sidelobe suppression processing on the image columns.

[0030] Optionally, processing the processed target slice image using a target detection algorithm to obtain a target binary map, including:

[0031] Determine a global detection threshold for detecting the ship target in the target slice image using a preset model algorithm;

[0032] Compare each pixel value in the target slice image with the global detection threshold, and determine whether the current pixel value is greater than the global detection threshold;

[0033] When the current pixel value is greater than the global detection threshold, determine that the current pixel value is a target point, and set the pixel value of the target point to 255;

[0034] When the current pixel value is not greater than the global detection threshold, determine that the current pixel value is a background point, and set the pixel value of the background point to 0.

[0035] In a second aspect, the present application discloses a target detection device based on high-resolution SAR, including:

[0036] A slice image acquisition module, configured to acquire a slice image containing a sea ship target based on high-resolution SAR, and determine the brightness mean of the slice image, the total row brightness value of each row in the slice image, and the total column brightness value of each column;

[0037] A scattered sidelobe suppression region determination module, configured to screen out the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values, and determine the boundary positions of the corresponding scattered sidelobe suppression regions based on the brightness mean, the brightness value of the target row, and the brightness value of the target column;

[0038] A scattered sidelobe suppression processing module, configured to determine a target slice image that needs to be subjected to scattered sidelobe suppression processing on the slice image based on preset conditions using the interval between the boundary positions, and process the processed target slice image using a target detection algorithm to obtain a target binary map;

[0039] A minimum circumscribed rectangle determination module, configured to determine the minimum circumscribed rectangle of the ship target based on the region where the ship target is located in the target binary map, so as to detect the ship target.

[0040] In a third aspect, the present application discloses an electronic device, which includes a processor and a memory; wherein, the memory is used to store a computer program, and the computer program is loaded and executed by the processor to implement the target detection method based on high-resolution SAR as described above.

[0041] In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program; wherein the computer program, when executed by a processor, implements the target detection method based on high-resolution SAR as described above.

[0042] In the present application, first, a sliced image containing sea surface ship targets based on high-resolution SAR is acquired, and the brightness mean value of the sliced image, the total row brightness value of each row in the sliced image, and the total column brightness value of each column in the sliced image are determined; then, the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values are selected, and based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column, the boundary positions of the corresponding scattering sidelobe suppression regions are determined; based on preset conditions, the interval between the boundary positions is used to determine the target sliced image that needs to be subjected to scattering sidelobe suppression processing on the sliced image, and the processed target sliced image is processed using a target detection algorithm to obtain a target binary image; finally, the minimum bounding rectangle of the ship target is determined according to the region where the ship target is located in the target binary image, so as to detect the ship target. It can be seen that by statistically calculating the total row and column brightness values of the sliced image of the high-resolution SAR sea surface ship target, searching for the "cross" or "one-word" trailing bright lines caused by the scattering sidelobes where the target may exist in the image, and then defining the boundary positions of the scattering sidelobe interference through the relationship between the image mean value and the scattering sidelobe intensity, without relying on prior knowledge such as manual parameter setting, fully considering the calculation efficiency and the problem that the strong scattering part of the target may be missearched and processed due to the reduction of the sidelobe scattering in the existing iterative search process, on the premise of ensuring automatic processing, it meets the accurate definition of the range of the "cross" or "one-word" trailing bright lines caused by the target scattering sidelobes in the sliced image of the high-resolution SAR sea surface ship target. In addition, the neighborhood region is used to process the target scattering sidelobe region. On this basis, the segmentation processing of the sea surface ship target and the accurate extraction of the target MBR are carried out to achieve the effect of suppressing and eliminating the target scattering sidelobe interference, realizing the adaptive extraction of the minimum bounding rectangle of the ship target, with the advantages of small calculation amount, fast processing speed, no manual intervention in the whole process, and relatively accurate extraction, which is close to the application requirements of the automatic processing of SAR sea surface ship targets. By using the target scattering sidelobe suppression based on the SAR original sliced image in the present application instead of the traditional sidelobe suppression processing based on the target segmentation binary image, it avoids the problem that the latter expands or erodes the real target region during the morphological processing, resulting in serious distortion of the target MBR extraction. Description of the Drawings

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on the provided drawings.

[0044] Figure 1 Schematic diagram of the trailing phenomenon caused by the scattering sidelobes of sea vessel targets in high-resolution SAR disclosed in this application;

[0045] Figure 2 Flowchart of a target detection method based on high-resolution SAR disclosed in this application;

[0046] Figure 3 Flow framework diagram for extracting the minimum circumscribed rectangle of a target based on suppressing the scattering sidelobes of sea vessel targets in high-resolution SAR disclosed in this application;

[0047] Figure 4 Schematic diagram of the definition result of the scattering sidelobe suppression area of sea vessel target slices and the comparison result of the extraction of the minimum circumscribed rectangle of the target based on high-resolution SAR disclosed in this application;

[0048] Figure 5 Flowchart of a specific target detection method based on high-resolution SAR disclosed in this application;

[0049] Figure 6 Structural schematic diagram of a target detection device based on high-resolution SAR disclosed in this application;

[0050] Figure 7 Structural diagram of an electronic device disclosed in this application. Detailed implementation manners

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0052] Currently, it is difficult to achieve the effect of sidelobe scattering elimination when extracting the MBR of SAR sea vessels. Moreover, if the method of iterative search is used to process sidelobe scattering, the calculation time is large. In the later stage of iterative processing, as the sidelobe scattering interference weakens, other strong scattering parts in the vessel target area will be mis-searched and processed. Or when using the method of morphological processing of the target segmentation binary image to eliminate sidelobe scattering interference, it will inevitably expand or erode the real target area at the same time, resulting in serious distortion of the target MBR.

[0053] Therefore, this application discloses a target detection scheme based on high-resolution SAR, which can quickly and accurately extract high-resolution SAR sea vessel targets that meet the requirements of rapid application, and the minimum bounding rectangle of the target in the case of strong scattering sidelobes.

[0054] An embodiment of the present invention discloses a target detection method based on high-resolution SAR. Refer to Figure 2 As shown, the method includes:

[0055] Step S11: Obtain a slice image based on high-resolution SAR containing sea ship targets, and determine the brightness mean value of the slice image, the total row brightness value of each row in the slice image, and the total column brightness value of each column in the slice image.

[0056] In the embodiment of this application, it is mainly based on the target detection of high-resolution SAR. The scene described in the SAR image is very complex. Therefore, in order to accurately extract the attribute features of the target, a slice image based on high-resolution SAR containing sea ship targets is obtained to identify the targets in the slice image and improve the identification performance of SAR targets in complex scenes. It can be understood that the target detection method based on high-resolution SAR of the present invention is not limited to detecting sea ship targets on the sea surface. When applied to land targets or airspace targets, the slice image containing sea ship targets in the present invention scheme can also be replaced by a slice image of land targets or airspace targets, which is not specifically limited here.

[0057] In the embodiment of this application, based on the slice image of high-resolution SAR sea ship targets, the brightness mean value of the slice image is statistically calculated, and the row and column brightness total values of the image are calculated by using the statistical methods of row by row and column by column.

[0058] Further, when determining the brightness mean value of the slice image, obtain the width information and height information of the slice image; use the width information and the height information to determine the total number of pixels of the slice image; based on the total number of pixels and the gray levels of all pixels in the slice image, determine the brightness mean value of the slice image.

[0059] Specifically, when the width and height information of the slice image containing sea ship targets is obtained, use the formula Statistically calculate the brightness mean value of the entire image; among them, is the grayscale mean value of the image, N is the total number of pixels in the image, N = W × H, H is the total number of rows in the image, and W is the total number of columns in the image; s is the grayscale level of the image, and the value range is 0 ≤ s ≤ 255; s ij is the grayscale level of the pixel in the i-th row and j-th column of the figure. That is, based on the width information and height information of the slice image, the total pixel value of the slice image is statistically calculated in a point-by-point cumulative manner, and then the brightness mean value of the image is obtained by taking the average.

[0060] In the embodiment of the present application, determining the total row brightness value of each row in the slice image and the total column brightness value of each column includes: performing row-by-row statistics on the slice image, and determining the total row brightness value of each row in the slice image based on the height information and the grayscale levels of the pixels in each row of the slice image; performing column-by-column statistics on the slice image, and determining the total column brightness value of each column in the slice image based on the width information and the grayscale levels of the pixels in each column of the slice image.

[0061] Specifically, using the formula statistically calculate the total brightness value of each image row of the entire image, and use the formula statistically calculate the total brightness value of each image column of the entire image; where s is the grayscale level of the image, and the value range is 0 ≤ s ≤ 255; s ij is the grayscale level of the pixel in the i-th row and j-th / column of the figure; H is the total number of rows in the image, C i is the total image brightness value of the i-th row; s ji is the grayscale level of the pixel in the j-th column and i-th / row of the figure, W is the total number of columns in the image, and R j is the total image brightness value of the j-th column. That is, in the direction of the slice image row, calculate the total pixel value of all pixels in each image row row by row in a cumulative manner; in the direction of the slice image column, calculate the total pixel value of all pixels in each image column row by row in a cumulative manner.

[0062] Step S12: Screen out the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values, and determine the boundary positions of the corresponding scattering sidelobe suppression regions based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column.

[0063] In the embodiment of the present application, based on the total row and column brightness values statistically calculated row by row and column by column, search for and locate the row with the maximum brightness and the column with the maximum brightness in the slice image containing the sea surface ship target. That is, in the total row brightness value C i , use the method of row-by-row search and comparison to determine the value C max with the maximum row brightness in the target slice image, and record the position of the image row where the maximum value is located; correspondingly, in the total column brightness value Rj Among them, by using the method of column-by-column search and comparison, the maximum value R of the column brightness in the target slice image is determined. max The position of the image column where the maximum value is located is recorded.

[0064] In the embodiment of the present application, after determining the target row with the maximum brightness in the total row brightness value and the target column with the maximum brightness in the total column brightness value, it indicates that there may be problems of cross or one-word trailing interference caused by strong target scattering side lobes in this row and / or this column. Therefore, further, based on the statistical average image brightness and the maximum values of the total brightness of the searched image rows and columns, with the target row and target column as the center, the boundary ranges of the scattering side lobe suppression regions are calculated and searched on both sides. In this way, the boundary range of the scattering side lobe interference can be defined and processed.

[0065] Step S13: Based on preset conditions, use the interval between the boundary positions to determine the target slice image that needs to be subjected to scattering side lobe suppression processing for the slice image, and process the processed target slice image using a target detection algorithm to obtain a target binary image.

[0066] In the embodiment of the present application, based on the boundary positions determined with the target row and target column as the center, row and column scattering side lobe suppression processing is carried out on the range of the scattering side lobe suppression region. It can be understood that when there is a cross or one-word trailing phenomenon in the slice image, the boundary positions of the range of the scattering side lobe suppression region can be determined, and then based on preset conditions, it can be further determined whether the target scattering side lobe exists and whether suppression processing is required.

[0067] In the embodiment of the present application, when it is considered that the target scattering side lobe exists and scattering side lobe suppression processing is required in the slice image, the target slice image is obtained, and then the target slice image is subjected to scattering side lobe suppression processing. Further, a target detection algorithm is used to process the target slice image after the scattering side lobe suppression processing, and a target binary image after the scattering side lobe suppression is obtained through the target detection processing.

[0068] Specifically, the processing of the processed target slice image using a target detection algorithm to obtain a target binary image includes: using a preset model algorithm to determine the global detection threshold for detecting the ship target in the target slice image; comparing each pixel value in the target slice image with the global detection threshold, and determining whether the current pixel value is greater than the global detection threshold; when the current pixel value is greater than the global detection threshold, determining that the current pixel value is a target point and setting the pixel value of the target point to 255; when the current pixel value is not greater than the global detection threshold, determining that the current pixel value is a background point and setting the pixel value of the background point to 0.

[0069] Exemplarily, considering that the sliced image is relatively small and the background is relatively uniform, the embodiments of the present application can adopt a global CFAR (Constant False Alarm Rate Detector) target detection algorithm to detect the target sliced image and obtain a target detection binary map. By using the global CFAR target detection algorithm and based on the pre-set constant false alarm rate p fa , through the calculation of the global statistical characteristics of the sliced image, and by using , the global target detection threshold x 0 is obtained; then, a point-by-point processing method is adopted to compare each pixel value in the image with the threshold. If it is greater than the threshold, it is judged as a target point, and the pixel value is set to 255; otherwise, it is judged as the background, and the pixel value is set to 0 until the processing of the entire image is completed to form a target segmentation binary map. Among them, p(x) is the probability density function of the SAR image clutter distribution model, and the Rayleigh distribution is selected.

[0070] It should be noted that the acquisition of the target binary map is not limited to the above CFAR algorithm. The maximum inter-class segmentation algorithm can also be adopted. By means of a sliding window calculation method, each pixel in the image is compared with the threshold. If it is greater than the threshold, it is judged as a target point, and the pixel value is set to 255; otherwise, it is judged as the background, and the pixel value is set to 0 until the processing of the entire image is completed to form a target segmentation binary map.

[0071] Step S14: Determine the minimum bounding rectangle of the ship target according to the area where the ship target is located in the target binary map, so as to detect the ship target.

[0072] In the embodiments of the present application, based on the ship target area in the target binary map, a mathematical processing method is used to calculate and mark the minimum bounding rectangle of the ship target. In this way, the extraction of the minimum bounding rectangle of the target for suppressing the scattering sidelobe of the high-resolution SAR sea surface ship target is realized.

[0073] As Figure 3 shown, it is a flowchart of the method for extracting the minimum bounding rectangle of the target for suppressing the scattering sidelobe of the high-resolution SAR sea surface ship target in the embodiments of the present application, and the order of its processing is from top to bottom. It should be understood that in the disclosed process, the specific order or level of the steps is an example of the exemplary method. Based on design preferences, it should be understood that the specific order or level of the steps in the process can be rearranged without departing from the protection scope of the present disclosure. The appended method claims give the elements of the various steps in an exemplary order and are not limited to a specific order or level.

[0074] As Figure 4The figure shows the comparison results of the defined scattering sidelobe suppression region and the extraction of the minimum circumscribed rectangle of the target for high-resolution SAR sea vessel targets according to the embodiments of the present invention. It can be seen from the figure that (a) is a SAR image slice containing ship targets; (b) shows the defined suppression region; (c) is the minimum circumscribed rectangle of the target extracted without suppression; (d) is the minimum circumscribed rectangle of the target extracted by the present solution. Obviously, due to the trailing phenomenon in (c), the extracted minimum circumscribed rectangle has deviated seriously from the MBR of the real target, while the method of the present solution can define and suppress the range of the scattering sidelobe trailing region, ultimately improving the accuracy of target MBR extraction.

[0075] In this application, first, a sliced image containing sea surface ship targets based on high-resolution SAR is obtained, and the brightness mean of the sliced image, the total row brightness value of each row in the sliced image, and the total column brightness value of each column in the sliced image are determined; then, the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values are selected, and based on the brightness mean, the brightness value of the target row, and the brightness value of the target column, the boundary positions of the corresponding scattering sidelobe suppression regions are determined; based on preset conditions, the target sliced image that needs to be subjected to scattering sidelobe suppression processing is determined by using the interval between the boundary positions, and the processed target sliced image is processed by using a target detection algorithm to obtain a target binary image; finally, the minimum circumscribed rectangle of the ship target is determined according to the region where the ship target is located in the target binary image, so as to detect the ship target. It can be seen that by statistically calculating the total row and column brightness values of the sliced image of the high-resolution SAR sea surface ship target, searching for the "cross" or "one-word" trailing bright lines caused by the scattering sidelobes where the target may exist in the image, and then defining the boundary positions of the scattering sidelobe interference through the relationship between the image mean and the scattering sidelobe intensity, it does not rely on prior knowledge such as manual parameter setting, fully considering the calculation efficiency and the problem that the strong scattering part of the target may be mis-searched and processed due to the reduction of the sidelobe scattering in the existing iterative search process. On the premise of ensuring automatic processing, it meets the accurate definition of the range of the "cross" or "one-word" trailing bright lines caused by the target scattering sidelobes in the sliced image of the high-resolution SAR sea surface ship target. In addition, the neighborhood region is used to process the target scattering sidelobe region. On this basis, the segmentation processing of the sea surface ship target and the accurate extraction of the target MBR are carried out, achieving the effect of suppressing and eliminating the target scattering sidelobe interference, realizing the adaptive extraction of the minimum circumscribed rectangle of the ship target, and having the advantages of small calculation amount, fast processing speed, no manual intervention in the whole process, and relatively accurate extraction, which is close to the application requirements of the automatic processing of the SAR sea surface ship target. By using the target scattering sidelobe suppression based on the SAR original sliced image in this application instead of the traditional sidelobe suppression processing based on the target segmentation binary image, it avoids the problem that the real target region is enlarged or eroded during the morphological processing of the latter, resulting in serious distortion of the target MBR extraction.

[0076] An embodiment of this application discloses a specific target detection method based on high-resolution SAR. Refer to Figure 5 as shown, this method includes:

[0077] Step S21: Obtain a sliced image containing sea surface ship targets based on high-resolution SAR, and determine the brightness mean of the sliced image, the total row brightness value of each row in the sliced image, and the total column brightness value of each column in the sliced image.

[0078] Among them, for the more specific processing process of the above step S21, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be elaborated herein.

[0079] Step S22: Screen out the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values, and construct a threshold relational expression for defining the range of the scattering sidelobe suppression region. Based on the threshold relational expression, use the brightness mean value and the brightness value of the target row to determine a first result threshold, and use the brightness mean value and the brightness value of the target column to determine a second result threshold.

[0080] In the embodiment of the present application, based on the total row and column brightness values and the image brightness mean value statistically calculated row by row and column by column, a threshold calculation formula for defining the range of the target scattering sidelobe suppression region is constructed. In a specific implementation manner, according to the maximum value C of the image row brightness max and the image mean value construct a threshold calculation formula for defining the boundary of the target scattering sidelobe suppression region in the image row direction In another implementation manner, according to the difference relationship between the maximum value R of the image column brightness max and the image mean value construct a threshold calculation formula for defining the boundary of the target scattering sidelobe suppression region in the image column direction

[0081] Step S23: Centering on the target row, compare the total row brightness values of other rows within a preset pixel range on both sides of the target row with the first result threshold. When the total row brightness value of the other rows is less than the first result threshold, determine that the position of the other rows is the boundary position of the scattering sidelobe suppression region.

[0082] In the embodiment of the present application, centering on the target row, search for the boundary of the scattering sidelobe suppression region within a certain range on both sides. Sequentially compare the total row brightness values of the statistically calculated image rows with the calculated threshold for defining the range of the target scattering sidelobe suppression region on both sides. When the search row is within the search range and its C i value is less than the first result threshold, determine that the current row is the boundary of the scattering sidelobe suppression region, otherwise it is considered that there is no suppression boundary.

[0083] Exemplarily, when the total row brightness values C of each image are statistically obtained i after that, centering on the recorded row C with the maximum brightness max within a range of 30 pixels on both sides, respectively compare the total brightness values C of the rows on both sides i with B in step S22 c When the search row offset does not reach 30 pixels and its C iWhen the value is less than the threshold, the row where it is located is determined as the boundary of the scattering sidelobe suppression region; otherwise, it is considered that there is no suppression boundary on this side, and its boundary position is directly set to C max+30 Finally, the boundary positions P C1 and P C2 (P C1 <P C2 ) of the row-direction scattering sidelobe suppression region on both sides of the row with the maximum brightness are obtained.

[0084] Step S24: Centering on the target column, compare the total column brightness value of other columns within the preset pixel range on both sides of the target column with the second result threshold. When the total column brightness value of the other columns is less than the second result threshold, determine the position of the other columns as the boundary position of the scattering sidelobe suppression region.

[0085] In the embodiment of the present application, centering on the target column, search for the boundary of the scattering sidelobe suppression region within a certain range on both sides, and sequentially compare the total column brightness value of the statistically obtained image columns with the calculated threshold for defining the range of the target scattering sidelobe suppression region on both sides. When the search row is within the search range and its R i value is less than the second result threshold, determine the column where it is located as the boundary of the scattering sidelobe suppression region; otherwise, it is considered that there is no suppression boundary.

[0086] Exemplarily, when the total row brightness value R i of each image is statistically obtained, within a range of 30 pixels on both sides of the recorded row with the maximum brightness R max as the center, respectively compare the total brightness value R i of each row on both sides with B R in step S22. When the offset of the search row does not reach 30 pixels and its R i value is less than the threshold, the column where it is located is determined as the boundary of the scattering sidelobe suppression region; otherwise, it is considered that there is no suppression boundary on this side, and its boundary position is directly set to R max+30 . Finally, the boundary positions P R1 and P R2 (P R1 <P R2 ) of the column-direction scattering sidelobe suppression region on both sides of the column with the maximum brightness are obtained.

[0087] Step S25: When the interval between the boundary positions is less than the preset pixel value, it is determined that there are scattering sidelobes between the boundary positions, and then perform scattering sidelobe suppression processing on the sliced image to obtain a target sliced image, and process the processed target sliced image using a target detection algorithm to obtain a target binary image.

[0088] In the embodiment of the present application, whether to perform target scattering sidelobe suppression processing depends on the interval size between the two boundary values obtained by searching in step S23 and step S24. When the interval is less than the preset pixel value, it is considered that the target scattering sidelobe exists, and then the suppression processing is performed; otherwise, no processing is performed.

[0089] In the embodiment of the present application, when performing scattering sidelobe suppression processing on the target slice image, the pixel value used to replace the image row in the scattering sidelobe suppression area is determined based on the brightness value of the target row and the boundary position corresponding to the target row, so as to perform the scattering sidelobe suppression processing on the image row; the pixel value used to replace the image column in the scattering sidelobe suppression area is determined based on the brightness value of the target column and the boundary position corresponding to the target column, so as to perform the scattering sidelobe suppression processing on the image column.

[0090] It should be noted that, based on the images in the left and right neighborhoods of the scattering sidelobe suppression area, a point-by-point processing method is adopted to sequentially calculate the mean values of the left and right adjacent pixels of each point in the suppression area within a certain neighborhood range, and use it to replace the pixel value of this point in the suppression area, so as to complete the scattering sidelobe suppression processing in the image row and column directions.

[0091] In a specific embodiment, based on P obtained by searching in step S23 C1 and P C2 , when the interval between P C2 and P C1 is less than 20 pixels, it is considered that the scattering sidelobe in the image row direction exists, and the following formula is used to perform the scattering sidelobe suppression processing in the image row direction:

[0092] Pos 1 =(C max -P C1 )+(C max -i)(P C1 <i<P C2 );

[0093] Pos 2 =(C max -P C2 )+(C max -i)(P C1 <i<P C2 );

[0094]

[0095] In another specific embodiment, based on P obtained by searching in step S33 R1 and P R2 , when the interval between P R1 and P R2When the interval is less than 20 pixels, it is considered that there are scattered sidelobes in the column direction of the image, and the scattered sidelobe suppression processing in the row direction of the image is performed using the formula:

[0096] Pos 1 =(R max -P R1 )+(R max -i)(P R1 <i<P R2 );

[0097] Pos 2 =(R max -P R2 )+(R max -i)(P R1 <i<P R2 );

[0098]

[0099] Step S26: Determine the minimum bounding rectangle of the ship target according to the area where the ship target is located in the target binary map, so as to detect the ship target.

[0100] Among them, for the more specific processing process of the above step S26, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be elaborated here.

[0101] It can be seen that by statistically calculating the total brightness values of the rows and columns of the sliced images of high-resolution SAR sea surface ship targets, searching for the "cross" or "one-line" trailing bright lines caused by the scattered sidelobes where the targets may exist in the images, and then defining the boundary positions of the scattered sidelobe interference through the relationship between the image mean value and the scattered sidelobe intensity, without relying on prior knowledge such as manual parameter setting, fully considering the calculation efficiency and the problem that the strong scattering parts of the targets may be mis-searched and processed due to the reduction of the scattered sidelobes during the existing iterative search process, under the premise of ensuring automatic processing, accurately defining the range of the "cross" or "one-line" trailing bright lines caused by the scattered sidelobes of the sea surface ship targets in the high-resolution SAR sliced images. In addition, the neighborhood area is used to process the target scattered sidelobe area. On this basis, the segmentation processing of the sea surface ship targets and the accurate extraction of the target MBR are carried out, achieving the effect of suppressing and eliminating the interference of the target scattered sidelobes, realizing the adaptive extraction of the minimum bounding rectangle of the ship targets, having the advantages of small calculation amount, fast processing speed, no manual intervention throughout the process and relatively accurate extraction, and being close to the application requirements of the automatic processing of SAR sea surface ship targets. By using the target scattered sidelobe suppression based on the SAR original sliced images in this application instead of the traditional sidelobe suppression processing based on the target segmentation binary map, it avoids the problem that the true target area is enlarged or eroded during the morphological processing in the latter, resulting in serious distortion of the target MBR extraction.

[0102] Correspondingly, the embodiment of the present application also discloses a target detection device based on high-resolution SAR. Refer to Figure 6 as shown, the device includes:

[0103] A slice image acquisition module 11, configured to acquire a slice image containing a sea surface ship target based on high-resolution SAR, and determine the brightness mean value of the slice image, the total row brightness value of each row in the slice image, and the total column brightness value of each column in the slice image;

[0104] A scattered sidelobe suppression region determination module 12, configured to screen out the target row with the largest brightness in the total row brightness values and the target column with the largest brightness in the total column brightness values, and determine the boundary positions of the corresponding scattered sidelobe suppression regions respectively based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column;

[0105] A scattered sidelobe suppression processing module 13, configured to determine a target slice image that needs to be subjected to scattered sidelobe suppression processing on the slice image based on a preset condition by using the interval between the boundary positions, and process the processed target slice image by using a target detection algorithm to obtain a target binary image;

[0106] A minimum circumscribed rectangle determination module 14, configured to determine the minimum circumscribed rectangle of the ship target according to the region where the ship target is located in the target binary image, so as to detect the ship target.

[0107] Among them, for the more specific working processes of the above-mentioned various modules, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details are not described herein again.

[0108] It can be seen that through the above solution of this embodiment, first, a sliced image containing sea surface ship targets based on high-resolution SAR is obtained, and the brightness mean value of the sliced image, the total row brightness value of each row in the sliced image, and the total column brightness value of each column in the sliced image are determined; then, the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values are selected, and the boundary positions of the respective corresponding scattering sidelobe suppression regions are determined based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column; based on preset conditions, the target sliced image that needs to be subjected to scattering sidelobe suppression processing is determined by using the interval between the boundary positions, and the processed target sliced image is processed by using a target detection algorithm to obtain a target binary map; finally, the minimum circumscribed rectangle of the ship target is determined according to the region where the ship target is located in the target binary map, so as to detect the ship target. It can be seen that by statistically calculating the total row and column brightness values of the sliced image of the high-resolution SAR sea surface ship target, searching for the "cross" or "one-word" trailing bright lines caused by the scattering sidelobes where the target may exist in the image, and then defining the boundary positions of the scattering sidelobe interference through the relationship between the image mean value and the scattering sidelobe intensity, it does not rely on prior knowledge such as manual parameter setting, fully considers the calculation efficiency and the problem that the strong scattering part of the target may be mis-searched and processed due to the reduction of the sidelobe scattering in the existing iterative search process, and on the premise of ensuring automatic processing, accurately defines the range of the "cross" or "one-word" trailing bright lines caused by the target scattering sidelobes in the sliced image of the high-resolution SAR sea surface ship target. In addition, the neighborhood region is used to process the target scattering sidelobe region. On this basis, the segmentation processing of the sea surface ship target and the accurate extraction of the target MBR are carried out, achieving the effect of suppressing and eliminating the target scattering sidelobe interference, realizing the adaptive extraction of the minimum circumscribed rectangle of the ship target, and having the advantages of small calculation amount, fast processing speed, no manual intervention in the whole process, and relatively accurate extraction, which is close to the application requirements of the automatic processing of the SAR sea surface ship target. By using the target scattering sidelobe suppression based on the SAR original sliced image in this application instead of the traditional sidelobe suppression processing based on the target segmentation binary map, it avoids the problem that the latter expands or erodes the real target region during the morphological processing, resulting in serious distortion of the target MBR extraction.

[0109] Furthermore, the embodiment of the present application also discloses an electronic device, Figure 7 which is the structure diagram of the electronic device 20 shown in an exemplary embodiment. The content in the figure cannot be regarded as any limitation on the use scope of the present application.

[0110] Figure 7Schematic diagram of the structure of an electronic device 20 provided by an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. Among them, the memory 22 is used to store a computer program, and the computer program is loaded and executed by the processor 21 to implement the relevant steps in the target detection method based on high-resolution SAR disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in this embodiment may specifically be a synthetic aperture radar.

[0111] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows is any communication protocol applicable to the technical solution of the present application, and no specific limitation is imposed on it here; the input / output interface 25 is used to obtain external input data or output data to the outside, and its specific interface type can be selected according to specific application needs, and no specific limitation is made here.

[0112] In addition, as a carrier for resource storage, the memory 22 may be a read-only memory, a random access memory, a magnetic disk, or an optical disc, etc. The resources stored thereon may include an operating system 221, a computer program 222, and data 223, etc. The data 223 may include various kinds of data. The storage method may be transient storage or permanent storage.

[0113] Among them, the operating system 221 is used to manage and control each hardware device on the electronic device 20 and the computer program 222, and it may be Windows Server, Netware, Unix, Linux, etc. In addition to the computer program capable of implementing the target detection method based on high-resolution SAR executed by the electronic device 20 disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0114] Furthermore, an embodiment of the present application also discloses a computer-readable storage medium. The computer-readable storage medium mentioned here includes random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, magnetic disks, or optical discs, or any other form of storage medium well-known in the technical field. Among them, when the computer program is executed by the processor, it implements the foregoing target detection method based on high-resolution SAR. For the specific steps of this method, reference may be made to the corresponding content disclosed in the foregoing embodiments, and details are not described herein again.

[0115] In the present specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description in the method section.

[0116] The steps of the target detection or algorithm based on high-resolution SAR described in combination with the embodiments disclosed herein can be implemented directly by hardware, software modules executed by a processor, or a combination of both. The software modules can be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium well-known in the technical field.

[0117] Finally, it should also be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the presence of another identical element in the process, method, article or device comprising the said element.

[0118] The above has introduced in detail the target detection method, device, equipment and storage medium based on high-resolution SAR provided by the present invention. Specific examples are used herein to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A target detection method based on high-resolution SAR, characterized in that, it includes: Obtain a sliced image based on high-resolution SAR containing sea surface ship targets, and determine the brightness mean value of the sliced image, as well as the total row brightness value of each row and the total column brightness value of each column in the sliced image; Select the target row with the maximum brightness in the total row brightness values and the target column with the maximum brightness in the total column brightness values, and determine the boundary positions of the corresponding scattering sidelobe suppression regions based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column; Based on preset conditions, use the interval between the boundary positions to determine the target sliced image that needs to be processed for scattering sidelobe suppression of the sliced image, and process the processed target sliced image using a target detection algorithm to obtain a target binary image; Determine the minimum bounding rectangle of the ship target according to the region where the ship target is located in the target binary image, so as to detect the ship target; Among them, determining the boundary positions of the corresponding scattering sidelobe suppression regions based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column includes: Construct a threshold relationship formula for defining the range of the scattering sidelobe suppression region; Based on the threshold relationship formula, use the brightness mean value and the brightness value of the target row to determine a first result threshold, and use the brightness mean value and the brightness value of the target column to determine a second result threshold; Centered on the target row, compare the total row brightness values of other rows within a preset pixel range on both sides of the target row with the first result threshold. When the total row brightness value of the other rows is less than the first result threshold, determine the position of the other rows as the boundary position of the scattering sidelobe suppression region; Centered on the target column, compare the total column brightness values of other columns within the preset pixel range on both sides of the target column with the second result threshold. When the total column brightness value of the other columns is less than the second result threshold, determine the position of the other columns as the boundary position of the scattering sidelobe suppression region.

2. The target detection method based on high-resolution SAR according to claim 1, characterized in that, determining the brightness mean value of the sliced image includes: Obtain the width information and height information of the sliced image; Use the width information and the height information to determine the total number of pixels of the sliced image; Based on the total number of pixels and the gray levels of all pixels in the sliced image, determine the brightness mean value of the sliced image.

3. The target detection method based on high-resolution SAR according to claim 2, characterized in that, determining the total row brightness value of each row and the total column brightness value of each column in the sliced image includes: Perform row-by-row statistics on the sliced image, and based on the height information and the gray levels of each row of pixels in the sliced image, determine the total row brightness value of each row in the sliced image; Perform column-by-column statistics on the sliced image, and determine the total column brightness value of each column in the sliced image based on the width information and the gray level of each column of pixels in the sliced image.

4. The method for target detection based on high-resolution SAR according to claim 1, wherein, based on the preset conditions, using the interval between the boundary positions to determine the target sliced image that needs to be subjected to scattering sidelobe suppression processing for the sliced image, includes: when the interval between the boundary positions is less than the preset pixel value, it is determined that there are scattering sidelobes between the boundary positions, and then the sliced image is subjected to scattering sidelobe suppression processing to obtain the target sliced image.

5. The method for target detection based on high-resolution SAR according to claim 1, wherein, after determining the target sliced image that needs to be subjected to scattering sidelobe suppression processing for the sliced image based on the preset conditions, it further includes: determining the pixel value for replacing the image row in the scattering sidelobe suppression area based on the brightness value of the target row and the boundary position corresponding to the target row, so as to perform the scattering sidelobe suppression processing on the image row; determining the pixel value for replacing the image column in the scattering sidelobe suppression area based on the brightness value of the target column and the boundary position corresponding to the target column, so as to perform the scattering sidelobe suppression processing on the image column.

6. The method for target detection based on high-resolution SAR according to any one of claims 1 to 5, wherein, processing the processed target sliced image using a target detection algorithm to obtain a target binary map, includes: using a preset model algorithm to determine the global detection threshold for detecting the ship target in the target sliced image; comparing each pixel value in the target sliced image with the global detection threshold, and determining whether the current pixel value is greater than the global detection threshold; when the current pixel value is greater than the global detection threshold, determining that the current pixel value is a target point, and setting the pixel value of the target point to 255; when the current pixel value is not greater than the global detection threshold, determining that the current pixel value is a background point, and setting the pixel value of the background point to 0.

7. A target detection device based on high-resolution SAR, wherein, includes: a sliced image acquisition module, configured to acquire a sliced image containing a sea ship target based on high-resolution SAR, and determine the brightness mean value of the sliced image, the total row brightness value of each row in the sliced image, and the total column brightness value of each column in the sliced image; a scattering sidelobe suppression area determination module, configured to screen out the target row with the maximum brightness in the total row brightness value and the target column with the maximum brightness in the total column brightness value, and determine the boundary positions of the respective corresponding scattering sidelobe suppression areas based on the brightness mean value, the brightness value of the target row, and the brightness value of the target column; The scattering sidelobe suppression processing module is used to determine a target slice image that needs to be subjected to scattering sidelobe suppression processing on the slice image based on preset conditions by using the interval between the boundary positions, and process the processed target slice image by using a target detection algorithm to obtain a target binary map; The minimum circumscribed rectangle determination module is used to determine the minimum circumscribed rectangle of the ship target according to the area where the ship target is located in the target binary map, so as to detect the ship target; Among them, the scattering sidelobe suppression area determination module is specifically used for: Construct a threshold relationship formula for defining the range of the scattering sidelobe suppression area; Based on the threshold relationship formula, use the brightness mean value and the brightness value of the target row to determine a first result threshold, and use the brightness mean value and the brightness value of the target column to determine a second result threshold; Centered on the target row, compare the total row brightness value of other rows within a preset pixel range on both sides of the target row with the first result threshold. When the total row brightness value of the other rows is less than the first result threshold, determine that the position of the other rows is the boundary position of the scattering sidelobe suppression area; Centered on the target column, compare the total column brightness value of other columns within the preset pixel range on both sides of the target column with the second result threshold. When the total column brightness value of the other columns is less than the second result threshold, determine that the position of the other columns is the boundary position of the scattering sidelobe suppression area.

8. An electronic device Characterized in that The electronic device includes a processor and a memory; wherein, the memory is used to store a computer program, and the computer program is loaded and executed by the processor to implement the target detection method based on high-resolution SAR according to any one of claims 1 to 6.

9. A computer-readable storage medium Characterized in that For storing a computer program; wherein the computer program, when executed by a processor, implements the target detection method based on high-resolution SAR according to any one of claims 1 to 6.

Images