Circular Synthetic Aperture Radar Moving Target Detection Method and Device

Abstract

This application relates to a method and device for moving target detection by circular synthetic aperture radar. The main technical solutions include: First, obtain circular synthetic aperture radar data, perform background difference processing on the circular synthetic aperture radar data to obtain an initial image sequence; then, perform morphological processing and clustering operation on the initial image sequence to obtain a clustered image sequence; perform target tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to target signals; finally, extract the feature values of each category in the tracking image sequence, calculate the cumulative rotation angle of each category in the tracking image sequence according to the feature values of each category; distinguish moving target signals and height fluctuation target signals according to the cumulative rotation angle of each category in the tracking image sequence, and delete the category corresponding to the height fluctuation target signal from the tracking image sequence. This application has the effect of improving the accuracy of moving target detection.

Description

Technical Field

[0001] This application relates to the technical field of moving target detection of circular synthetic aperture radar, and particularly to a method and device for moving target detection of circular synthetic aperture radar. Background Art

[0002] Circular Synthetic Aperture Radar (Circular-SAR) not only provides high-resolution and high-frame-rate SAR images, but also can detect a specific area for a long time and from multiple angles. Therefore, it is widely used in the task of detecting moving targets.

[0003] In traditional implementation methods, a single-channel circular SAR moving target detection method based on the logarithmic background difference method is generally adopted. This method uses the movement of target signals in the sub-aperture image sequence to detect moving targets. However, due to the curved trajectory of circular SAR, height-varying targets (such as towers, buildings, etc.) form circular defocusing during imaging and move in the image sequence, making the detection algorithm unable to effectively distinguish high-rise buildings and moving target signals, resulting in a relatively high false alarm rate for circular SAR moving target detection. Based on this situation, a clutter suppression technique of triple-view interference cancellation is also adopted. Change detection is performed based on the statistical analysis of the image after clutter suppression and the reference image, and constant false alarm and morphological processing are used to further reduce false alarms; or autofocus processing is used to improve the negative impact brought by moving target defocusing, improve the signal-to-clutter ratio, and detect moving targets and eliminate false alarms respectively through two-stage cell-averaging constant false alarm processing with different thresholds.

[0004] However, the above two moving target detection methods with false alarm suppression effects cannot effectively remove the defocusing formed by height-varying targets during imaging, resulting in the problem that they still move in the image sequence and cause false alarms. Summary of the Invention

[0005] Based on this, this application provides a method, device, equipment and storage medium for moving target detection of circular synthetic aperture radar to distinguish moving target signals and height-varying target signals in the image sequence, and remove false alarms generated by height-varying target signals, solving the problem of high false alarm rate caused by height-varying targets in the process of moving target detection of circular synthetic aperture radar, and achieving the effect of improving the accuracy of moving target detection.

[0006] In a first aspect, a method for moving target detection of circular synthetic aperture radar is provided. The method includes:

[0007] Obtain circular synthetic aperture radar data, perform background difference processing on the circular synthetic aperture radar data to obtain an initial image sequence;

[0008] Perform morphological processing and clustering operation on the initial image sequence to obtain a clustering image sequence;

[0009] Perform object tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories that correspond one-to-one with the target signals;

[0010] Extract the eigenvalue of each category in the tracking image sequence, and calculate the cumulative rotation angle of each category in the tracking image sequence according to the eigenvalue of each category;

[0011] Distinguish the moving target signal and the highly fluctuating target signal according to the cumulative rotation angle of each category in the tracking image sequence, and delete the category corresponding to the highly fluctuating target signal from the tracking image sequence.

[0012] According to an implementable manner in the embodiment of the present application, performing object tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories that correspond one-to-one with the target signals includes:

[0013] Obtain the category with the highest overlapping pixels in every two adjacent frames of images of the same target signal in the clustered image sequence;

[0014] If the number of categories with the highest overlapping pixels is equal to 1, use the prefabricated cross-ratio algorithm to calculate the pixel cross-ratio of every two adjacent frames of images;

[0015] Compare the pixel cross-ratio with the preset cross-ratio threshold. If the pixel cross-ratio is greater than the preset cross-ratio threshold, change the category of the overlapping pixels in the latter frame of the image to the category of the overlapping pixels in the former frame of the image;

[0016] If the pixel cross-ratio is less than or equal to the preset cross-ratio threshold, set the category of the overlapping pixels in the latter frame of the image to be empty, and so on, to obtain a tracking image sequence with multiple categories that correspond one-to-one with the target signals.

[0017] According to an implementable manner in the embodiment of the present application, performing object tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories that correspond one-to-one with the target signals includes:

[0018] Obtain the category with the highest overlapping pixels in every two adjacent frames of images of the same target signal in the clustered image sequence;

[0019] If the number of categories with the highest overlapping pixels is greater than 1, use the prefabricated cross-ratio algorithm to calculate the pixel cross-ratio of every two adjacent frames of images in each category with the highest overlapping pixels in the same target signal respectively;

[0020] Compare the calculated pixel cross-ratios to obtain the maximum pixel cross-ratio;

[0021] Compare the maximum pixel cross-ratio with a preset cross-ratio threshold. If the maximum pixel cross-ratio is greater than the preset cross-ratio threshold, change the category of the overlapping pixels in the subsequent frame image to the category of the overlapping pixels in the previous frame image;

[0022] If the maximum pixel cross-ratio is less than or equal to the preset cross-ratio threshold, set the category of the overlapping pixels in the subsequent frame image to be empty, and so on, to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals.

[0023] According to an implementable manner in the embodiments of the present application, extract the feature values of each category in the tracking image sequence, and calculate the cumulative rotation angle of each category in the tracking image sequence based on the feature values of each category, including:

[0024] Extract the feature values of each frame image of each category in the tracking image sequence based on the feature extraction function, and calculate the endpoint coordinate values of each frame image of each category according to the feature values of each category in the tracking image sequence;

[0025] Calculate the feature angle of each frame image according to the endpoint coordinate values of each frame image of each category in the tracking image sequence;

[0026] Subtract the feature angle of the first frame image of each category from the feature angle of the last frame image of each category in the tracking image sequence to obtain the cumulative rotation angle of each category in the tracking image sequence.

[0027] According to an implementable manner in the embodiments of the present application, distinguish the moving target signal and the height fluctuation target signal according to the cumulative rotation angle of each category in the tracking image sequence, and delete the category corresponding to the height fluctuation target signal from the tracking image sequence, including:

[0028] Compare the cumulative rotation angles of each category in the tracking image sequence with a preset rotation angle respectively;

[0029] When the cumulative rotation angle is greater than the preset rotation angle, determine that the category corresponding to the cumulative rotation angle is the height fluctuation target signal, and delete the category from the tracking image sequence;

[0030] If the cumulative rotation angle is less than or equal to the preset rotation angle, determine that the category corresponding to the cumulative rotation angle is the moving target signal and do not perform any processing.

[0031] According to an implementable manner in the embodiments of the present application, the method further includes:

[0032] According to the endpoint coordinate values of each frame image, obtain the center point coordinate values of each category in the tracking image sequence;

[0033] According to the endpoint coordinate values of each frame image of each category and the corresponding center point coordinate values, obtain the standard deviation of each category in the tracking image sequence;

[0034] Determine the moving target signal and the height fluctuation target signal according to the standard deviation of each category in the tracking image sequence, and delete the category corresponding to the height fluctuation target signal from the tracking image sequence.

[0035] According to an achievable manner in the embodiments of the present application, determining the moving target signal and the height fluctuation target signal according to the standard deviation of each category in the tracking image sequence, and deleting the category corresponding to the height fluctuation target signal from the tracking image sequence includes:

[0036] Compare the standard deviation of each category in the tracking image sequence with a preset standard deviation respectively;

[0037] When the standard deviation is less than the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is greater than the preset rotation angle, determine that the category corresponding to the standard deviation is the height fluctuation target signal, and delete the category from the tracking image sequence;

[0038] When the standard deviation is greater than or equal to the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is less than or equal to the preset rotation angle, determine that the category corresponding to the standard deviation is the moving target signal and do not perform any processing.

[0039] In a second aspect, a moving target detection device for a circular synthetic aperture radar is provided. The device includes:

[0040] A background difference processing module, configured to obtain circular synthetic aperture radar data, perform background difference processing on the circular synthetic aperture radar data to obtain an initial image sequence;

[0041] A morphology and clustering processing module, configured to perform morphology processing and clustering operation on the initial image sequence to obtain a clustered image sequence;

[0042] A target tracking module, configured to perform target tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories that correspond one-to-one to the target signals;

[0043] A feature extraction module, configured to extract the feature values of each category in the tracking image sequence, and calculate the cumulative rotation angle of each category in the tracking image sequence according to the feature values of each category;

[0044] A target discrimination module, configured to distinguish the moving target signal and the height fluctuation target signal according to the cumulative rotation angle of each category in the tracking image sequence, and delete the category corresponding to the height fluctuation target signal from the tracking image sequence.

[0045] In a third aspect, a computer device is provided, including:

[0046] At least one processor; and

[0047] A memory communicatively connected to at least one processor; wherein,

[0048] The memory stores computer instructions executable by at least one processor, and the computer instructions are executed by at least one processor to enable at least one processor to execute the method involved in the above first aspect.

[0049] In a fourth aspect, a computer-readable storage medium is provided, on which computer instructions are stored, characterized in that the computer instructions are used to cause a computer to execute the method involved in the above first aspect.

[0050] According to the technical content provided by the embodiments of the present application, first, circular synthetic aperture radar data is acquired, and the circular synthetic aperture radar data is subjected to background difference processing to obtain an initial image sequence; then, the initial image sequence is subjected to morphological processing and clustering operation to obtain a clustered image sequence; target tracking is performed on the clustered image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to target signals; finally, the feature values of each category in the tracking image sequence are extracted, and the cumulative rotation angle of each category in the tracking image sequence is calculated based on the feature values of each category; the moving target signal and the height-undulating target signal are distinguished according to the cumulative rotation angle of each category in the tracking image sequence, and the category corresponding to the height-undulating target signal is deleted from the tracking image sequence. Through the above operations of distinguishing the moving target signal and the height-undulating target signal and deleting the category corresponding to the height-undulating target from the tracking image sequence, while ensuring the detection of moving targets, the problem of high false alarm rate caused by defocusing of the height-undulating target imaging and resulting in false alarms is solved, and the effect of improving the accuracy of moving target detection by circular synthetic aperture radar is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is an application environment diagram of a moving target detection method for circular synthetic aperture radar in an embodiment;

[0052] Figure 2 It is a flowchart of a moving target detection method for circular synthetic aperture radar in an embodiment;

[0053] Figure 3 It is a schematic diagram of an erosion operation of a moving target detection method for circular synthetic aperture radar in an embodiment;

[0054] Figure 4 It is a schematic diagram of a dilation operation of a moving target detection method for circular synthetic aperture radar in an embodiment;

[0055] Figure 5 It is a flowchart of step 205 in a moving target detection method for circular synthetic aperture radar in an embodiment Figure 1 ;

[0056] Figure 6 Schematic diagram of step 205 in a method for moving target detection by a circular synthetic aperture radar in an embodiment Figure 2 ;

[0057] Figure 7 Schematic diagram of feature extraction in a method for moving target detection by a circular synthetic aperture radar in an embodiment;

[0058] Figure 8 Schematic diagram of calculating endpoint coordinate values in a method for moving target detection by a circular synthetic aperture radar in an embodiment;

[0059] Figure 9 Schematic diagram of calculating feature angles in a method for moving target detection by a circular synthetic aperture radar in an embodiment;

[0060] Figure 10 Schematic diagram of the motion trajectories of moving target signals and height fluctuation target signals in a method for moving target detection by a circular synthetic aperture radar in an embodiment;

[0061] Figure 11 Schematic diagram of the preferred process in a method for moving target detection by a circular synthetic aperture radar in an embodiment;

[0062] Figure 12 Structure block diagram of a moving target detection device for a synthetic aperture radar in an embodiment;

[0063] Figure 13 Schematic structural diagram of a computer device in an embodiment. Specific embodiments

[0064] The following further describes the present application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0065] For the convenience of understanding, the system to which the present application is applicable is first described. A method for moving target detection by a circular synthetic aperture radar provided by the present application can be applied to a system architecture as shown in Figure 1 The system includes: a terminal device 102 - network - server 104, and the terminal 102 communicates with the server 104 through the network. Among them, the terminal 102 can be, but is not limited to, various hardware devices such as personal computers, laptop computers, and tablet computers, and the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0066] Figure 2 Flowchart of a method for moving target detection by a circular synthetic aperture radar provided by an embodiment of the present application. This method can be implemented by a device as shown in Figure 1executed by server 104 in the system architecture shown. As Figure 2 shown, the method may include the following steps:

[0067] Step S201: Obtain circumferential synthetic aperture radar data, perform background difference processing on the circumferential synthetic aperture radar data, and obtain an initial image sequence.

[0068] Here, after obtaining the circumferential synthetic aperture radar data, the server can perform background difference processing on the circumferential synthetic aperture radar data.

[0069] Among them, background difference processing is to perform difference between the background image and the foreground image to obtain an initial image sequence. Specifically, the manifestation form of the circumferential synthetic aperture radar data can be a sub-aperture image sequence. First, perform radiometric correction on the sub-aperture image sequence to obtain a radiometric image; then copy the radiometric image to obtain a copied radiometric image, perform median filtering on the copied radiometric image to obtain a background image; finally, perform difference processing between the background image and the radiometric image to obtain the foreground image, that is, the initial image sequence. Through the operation of background difference processing, it is convenient to detect and identify moving targets.

[0070] Step S203: Perform morphological processing and clustering operation on the initial image sequence to obtain a clustered image sequence.

[0071] Among them, morphological processing may include erosion processing and dilation processing.

[0072] As Figure 3 shown, first select a corresponding erosion template according to the morphology of the initial image sequence, and its erosion expression is as follows:

[0073]

[0074] Among them, "-" is the operator of erosion; A represents the initial image sequence; B represents the erosion template.

[0075] By performing erosion operation on the initial image sequence A and the erosion template B, scanning each pixel point of the initial image sequence A, performing "AND" operation on the pixel points of the initial image sequence A and the pixel points of the erosion template B, obtaining the minimum value of the pixel points in the covered area of the erosion template B, and using this minimum value to replace the pixel value of the reference point of the initial image sequence A. That is, after the original image sequence A is subjected to erosion processing, an erosion image sequence A - B, that is, A1, is obtained. It can be seen from the erosion image sequence A1 that after the erosion operation, the black high-bright points in the original image sequence A become fewer, realizing the erosion of the residual clutter in the original image sequence A.

[0076] As Figure 4As shown, after obtaining the corrosion image sequence A1, dilation processing is still required. According to the morphology of the corrosion image sequence A1, the corresponding dilation template B1 is selected, and the expression of its dilation processing is as follows:

[0077]

[0078] Among them, represents the operator of dilation; A1 represents the corrosion image sequence; B1 represents the dilation template.

[0079] Specifically, by performing dilation processing on the corrosion image sequence A1, scanning each pixel point in the corrosion image sequence A1, performing an "OR" operation on the pixel points of the dilation template B1 and the pixel points of the corrosion image sequence A1, obtaining the maximum value of the pixel points in the covered area of the dilation template B1, and replacing the pixel value of the reference point of the corrosion image sequence A1 with this maximum value to achieve dilation processing. That is, after dilating the corrosion image sequence A1, the dilated image sequence A1 + B1, namely A2, is obtained. After the dilation operation, the black high-bright points in the corrosion image sequence A1 become more, realizing the repair of the fracture area.

[0080] Here, after subjecting the original image sequence to morphological processing, a dilated image sequence is obtained, and further clustering operations are performed on the dilated image sequence.

[0081] Specifically, there are two important setting parameters for the clustering operation: the neighborhood radius and the number of samples in the neighborhood, which can be used to reflect the tightness of the sample distribution in the neighborhood. Arbitrarily select a pixel point without a category in the dilated image sequence as a seed, and select a set of pixel points that are density-reachable by this seed object, which is a clustering category cluster; then, select another pixel point without a category as a seed, and select a set of pixel points that are density-reachable by this seed object, which is another clustering category cluster; and so on, until all pixel points have categories, that is, a clustered image sequence is obtained.

[0082] Step S205, perform object tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals.

[0083] Here, by performing object tracking on the clustered image sequence, a tracking image sequence with multiple categories corresponding one-to-one to the target signals is obtained. It should be emphasized that the categories in the tracking image sequence are the above-mentioned clustering category clusters, each clustering category cluster is a category, and each category represents a different target signal.

[0084] Step S207, extract the feature values of each category in the tracking image sequence, and calculate the cumulative rotation angles of each category in the tracking image sequence based on the feature values of each category.

[0085] Specifically, the eigenvalue of each category in the tracking image sequence is extracted, and the corresponding cumulative rotation angle is calculated based on the eigenvalue of each category. It should be noted that the reason for obtaining the cumulative rotation angle is that the difference in the cumulative rotation angles of the moving target signal and the height fluctuation target signal is large, which is convenient for distinguishing between the two.

[0086] Step S209: Distinguish the moving target signal and the height fluctuation target signal according to the cumulative rotation angles of each category in the tracking image sequence, and delete the category corresponding to the height fluctuation target signal from the tracking image sequence.

[0087] Specifically, the moving target signal and the height fluctuation target signal are distinguished according to the cumulative rotation angles of each category in the tracking image sequence. After determining that it is a height fluctuation target signal, the category corresponding to the height fluctuation target signal is deleted from the tracking image sequence to achieve the effect of reducing the false alarm rate.

[0088] It can be seen that in the embodiment of the present application, first, the circular synthetic aperture radar data is obtained, and the circular synthetic aperture radar data is subjected to background difference processing to obtain an initial image sequence; then, the initial image sequence is subjected to morphological processing and clustering operation to obtain a clustering image sequence; target tracking is performed on the clustering image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals; finally, the eigenvalue of each category in the tracking image sequence is extracted, and the cumulative rotation angle of each category in the tracking image sequence is calculated according to the eigenvalue of each category; the moving target signal and the height fluctuation target signal are distinguished according to the cumulative rotation angles of each category in the tracking image sequence, and the category corresponding to the height fluctuation target signal is deleted from the tracking image sequence. Through the above operations of distinguishing the moving target signal and the height fluctuation target signal and deleting the category corresponding to the height fluctuation target from the tracking image sequence, while ensuring the detection of the moving target, the problem of high false alarm rate caused by defocusing of the height fluctuation target imaging is solved, and the effect of improving the accuracy of moving target detection by the circular synthetic aperture radar is achieved.

[0089] Refer to Figure 5 , in some embodiments, the above step 205 "perform target tracking on the clustering image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals" may specifically include the following steps.

[0090] Step S301: Obtain the category with the highest overlapping pixels in every two adjacent frames of the same target signal in the clustering image sequence.

[0091] Here, since the categories after clustering in each frame of the clustering image sequence are inconsistent, the same target signal may be clustered into different categories. Therefore, the category with the highest overlapping pixels in every two adjacent frames of the same target signal in the clustering image sequence can be obtained for unification.

[0092] In step S303, if the number of categories with the highest overlapping pixels is equal to 1, the prefabricated cross-ratio algorithm is used to calculate the pixel cross-ratio of every two adjacent frames of images of the same target signal.

[0093] Here, when the number of categories with the highest overlapping pixels is equal to 1, the prefabricated cross-ratio algorithm can be directly used to calculate the pixel cross-ratio of every two adjacent frames of images of the same target signal.

[0094] Specifically, the expression of the prefabricated cross-ratio algorithm is as follows:

[0095]

[0096] where B i , B i+1 are respectively the number of pixel points occupied by the same target signal in the i-th and (i + 1)-th frames of images. The numerator on the right side of the equal sign represents the number of pixel points in the overlapping area of the i-th and (i + 1)-th frames of images, and the denominator represents the total number of pixel points in the i-th and (i + 1)-th frames of images. Through this expression, the pixel cross-ratio of two adjacent frames of images of the same target signal can be obtained.

[0097] In step S305, the pixel cross-ratio is compared with a preset cross-ratio threshold. If the pixel cross-ratio is greater than the preset cross-ratio threshold, the category of the overlapping pixels in the latter frame of the image is changed to the category of the overlapping pixels in the former frame of the image.

[0098] Among them, the preset cross-ratio threshold can be set between 0.3 and 0.5.

[0099] Here, the pixel cross-ratio is compared with the preset cross-ratio threshold. If the pixel cross-ratio is greater than the preset cross-ratio threshold, the category of the overlapping pixels in the latter frame of the image of the same target signal is changed to the category of the overlapping pixels in the former frame of the image to achieve the unification of the categories of the same target signal.

[0100] In step S307, if the pixel cross-ratio is less than or equal to the preset cross-ratio threshold, the category of the overlapping pixels in the latter frame of the image is set to empty, and so on, to obtain a tracking image sequence with multiple categories corresponding one by one to the target signals.

[0101] Here, if the pixel cross-ratio is less than or equal to the preset cross-ratio threshold, the category of the overlapping pixels in the latter frame of the image is set to empty, and so on, until the categories of all target signals in the clustering image sequence are unified, and the tracking is completed to obtain a tracking image sequence with multiple categories corresponding one by one to the target signals.

[0102] Refer to Figure 6, in some other embodiments, the above step 205, "performing object tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals", may specifically include the following steps.

[0103] Step S401: Obtain the category with the highest overlapping pixels in every two adjacent frames of the same target signal in the clustered category image sequence.

[0104] Here, similar to the above step S301, it is still necessary to obtain the category with the highest overlapping pixels in every two adjacent frames of the same target signal in the clustered image sequence for further unification.

[0105] Step S403: If the number of categories with the highest overlapping pixels is greater than 1, use the prefabricated cross-ratio algorithm to calculate the pixel cross-ratio of every two adjacent frames in each category with the highest overlapping pixels of the same target signal respectively.

[0106] Here, since the number of categories with the highest overlapping pixels is greater than 1, it is necessary to use the prefabricated cross-ratio algorithm to calculate the pixel cross-ratio of every two adjacent frames in each category with the highest overlapping pixels of the same target signal respectively. The expression of the prefabricated cross-ratio algorithm is the same as above and will not be elaborated here.

[0107] Step S405: Compare the calculated pixel cross-ratios to obtain the maximum pixel cross-ratio.

[0108] Here, since the number of categories with the highest overlapping pixels is greater than 1, multiple pixel cross-ratios can be obtained. Compare the pixel cross-ratios to obtain the maximum pixel cross-ratio for comparison with the preset cross-ratio threshold.

[0109] Step S407: Compare the maximum pixel cross-ratio with the preset cross-ratio threshold. If the maximum pixel cross-ratio is greater than the preset cross-ratio threshold, change the category of the overlapping pixels in the subsequent frame image to the category of the overlapping pixels in the previous frame image.

[0110] Here, compare the maximum pixel cross-ratio with the preset cross-ratio threshold. If the maximum pixel cross-ratio is greater than the preset cross-ratio threshold, change the category of the overlapping pixels in the subsequent frame image of the same target signal to the category of the overlapping pixels in the previous frame image to achieve the unification of the categories of the same target signal.

[0111] Step S409: If the maximum pixel cross-ratio is less than or equal to the preset cross-ratio threshold, set the category of the overlapping pixels in the subsequent frame image to be empty, and so on, to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals.

[0112] If the maximum pixel cross-ratio is less than or equal to a preset cross-ratio threshold, the category of the overlapping pixels in the subsequent frame image is set to empty, and so on, until the categories of all target signals in the clustered image sequence are unified. After the tracking is completed, a tracking image sequence with multiple categories corresponding one-to-one to the target signals is obtained.

[0113] In some embodiments, based on a feature extraction function, the feature values of each frame image of each category in the tracking image sequence are extracted, and the endpoint coordinate values of each frame image of each category are calculated according to the feature values of each category in the tracking image sequence; the feature angle of each frame image is calculated according to the endpoint coordinate values of each frame image of each category in the tracking image sequence; the difference between the feature angles of the first frame image and the last frame image of each category in the tracking image sequence is obtained to get the cumulative rotation angle of each category in the tracking image sequence.

[0114] Here, the expression of the feature extraction function is:

[0115] M = regionprops(L)

[0116] where M represents the feature value of each frame image; L represents the tracking image sequence; regionprops represents the feature extraction function.

[0117] Specifically, by inputting the tracking image sequence into the feature extraction function, the centroid feature value, equivalent ellipse feature value, major axis feature value of the equivalent ellipse, and the included angle feature value between the major axis of the equivalent ellipse and the x-axis in the rectangular coordinate system of each frame image of each category in the tracking image sequence can be extracted. As Figure 7 shown, the point in Figure (A) is the extracted centroid feature value, the ellipse in Figure (B) is the extracted equivalent ellipse feature value, the dashed line in Figure (C) is the major axis feature value of the extracted equivalent ellipse, and the β angle in Figure (D) is the included angle feature value between the major axis of the effective ellipse and the x-axis in the rectangular coordinate system.

[0118] Referring to Figure 8 , the endpoint coordinate values of each frame image of each category are calculated according to the feature values of each category in the tracking image sequence. Here, assuming that the coordinates of the extracted centroid feature value are (X1, Y1), the major axis feature value of the equivalent ellipse is R M and the included angle feature value between the major axis of the equivalent ellipse and the x-axis in the rectangular coordinate system is β, the expression of the corresponding endpoint coordinate value of the equivalent ellipse is as follows:

[0119] P X1 = X1 - R M / 2 × cosβ

[0120] P Y1 = Y1 - R M / 2 × sinβ

[0121] Through the above calculations, the endpoint coordinate values P (P X1 , P Y1 ) of each frame image of each category in the tracking image sequence are obtained. According to the endpoint coordinate values of each frame image of each category in the tracking image sequence, the characteristic angle of each frame image is calculated. Here, considering the error of moving target detection, the endpoint coordinate values of the equivalent ellipse are definitely not in the same position, and they can be normalized and all moved to the origin, so that the coordinates of the new center-of-gravity eigenvalue are (X2, Y2). Referring to Figure 9 , the specific expression is as follows:

[0122] X2 = X1 - P X1

[0123] Y2 = Y1 - P Y1

[0124] According to the coordinates (X2, Y2) of the new center-of-gravity eigenvalue, the characteristic angle of each frame image can be obtained as:

[0125] tanθ1 = Y2 / X2

[0126] The difference between the characteristic angle of the first frame image and the characteristic angle of the last frame image of each category in the tracking image sequence is calculated, and its expression is as follows:

[0127] θ = θX - θ1

[0128] where θ1 is the characteristic angle of the first frame image; θX is the characteristic angle of the last frame image; through the above operations, the cumulative rotation angles of each category in the tracking image sequence are obtained.

[0129] In some embodiments, the cumulative rotation angles of each category in the tracking image sequence are respectively compared with a preset rotation angle; when the cumulative rotation angle is greater than the preset rotation angle, it is determined that the category corresponding to the cumulative rotation angle is a height fluctuation target signal, and the category is deleted from the tracking image sequence; if the cumulative rotation angle is less than or equal to the preset rotation angle, it is determined that the category corresponding to the cumulative rotation angle is a moving target signal and no processing is performed.

[0130] where the preset rotation angle can be set to about 60°, and this finally needs to be determined according to the specific variation law of the eigenvalue in the actual detection process.

[0131] Here, the cumulative rotation angles of each category in the tracking image sequence are respectively compared with a preset rotation angle; when the cumulative rotation angle is greater than the preset rotation angle, it is determined that the category corresponding to the cumulative rotation angle is a height fluctuation target signal, and the category is deleted from the tracking image sequence to achieve the effect of removing false alarms; if the cumulative rotation angle is less than or equal to the preset rotation angle, it is determined that the category corresponding to the cumulative rotation angle is a moving target signal and no processing is required.

[0132] In some embodiments, according to the endpoint coordinate values of each frame of image, the center point coordinate values of each category in the tracking image sequence are obtained; according to the endpoint coordinate values and the corresponding center point coordinate values of each frame of image of each category, the standard deviation of each category in the tracking image sequence is obtained, and the moving target signal and the height fluctuation target signal are determined according to the standard deviation of each category in the tracking image sequence, and the category corresponding to the height fluctuation target signal is deleted from the tracking image sequence.

[0133] Here, according to the endpoint coordinate values of each frame of image, the center point coordinate values of each category in the tracking image sequence are obtained, and the specific expression is as follows:

[0134] P XZ =(P X1 +P X2 +…+P XN ) / N

[0135] P YZ =(P Y1 +P Y2 +…+P YN ) / N

[0136] Among them, (P X1 , P Y1 ) represents the endpoint coordinate values of the first frame of image of a certain category; (P X2 , P Y2 ) represents the endpoint coordinate values of the second frame of image of this category; and so on, (P XN , P YN ) represents the endpoint coordinate values of the first frame of image of a certain category; (P XZ , P YZ ) represents the center point coordinate values of this category.

[0137] According to the endpoint coordinate values and the corresponding center point coordinate values of each frame of image of each category, the standard deviation of each category in the tracking image sequence is obtained. Here, it is necessary to first obtain the distance between the endpoint coordinate values and the corresponding center point coordinate values of each frame of image, and the specific expression is as follows:

[0138] P L1 =sqrt((P X1 -P XZ )^2+(P Y1 -P YZ )^2)

[0139] P L2 =sqrt((P X2 -P XZ )^2+(P Y2 -P YZ )^2)

[0140] And so on, P can be obtained. L3 、P L4 …P LN Then, by averaging the values of P L1 、P L2 …P LN The expression is as follows:

[0141] P LP =(P L1 +P L2 +…+P LN ) / N

[0142] Through the above expression, the average distance P LP of a certain category is obtained.

[0143] According to the distance between the endpoint coordinate value and the corresponding center point coordinate value of each frame of image and the average distance, the corresponding standard deviation is obtained. The specific expression is as follows:

[0144] S=sqrt(((P L1 -P LP )^2+(P L2 -P LP )^2+…(P LN -P LP )^2) / (N - 1))

[0145] Referring to Figure 10 , since the moving target signal has a wide range of motion in the tracking image sequence, the points roughly in a circular surround in the figure are the endpoints of the equivalent ellipse, and the central point is the obtained center point equivalent to the center of the circle. As can be seen from Figure A, the points in a circular surround are moving around the center point with a radius of r and a small range of motion; while from Figure B, it can be seen that the endpoints of the equivalent ellipse have a large range of motion, not a small-range rotational motion, but a linear-like motion. Therefore, the moving target signal and the highly fluctuating target signal can also be distinguished by calculating the standard deviation. That is, after obtaining the standard deviations of each category in the tracking image sequence, the moving target signal and the highly fluctuating target signal in the tracking image sequence can be determined according to the standard deviation. And after determining the highly fluctuating target signal, the category corresponding to the highly fluctuating target signal is deleted from the tracking image sequence.

[0146] In some embodiments, the standard deviations of each category in the tracking image sequence are respectively compared with a preset standard deviation; when the standard deviation is less than the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is greater than the preset rotation angle, then the category corresponding to the standard deviation is determined as the highly fluctuating target signal and the category is deleted from the tracking image sequence; when the standard deviation is greater than or equal to the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is less than or equal to the preset rotation angle, then the category corresponding to the standard deviation is determined as the moving target signal and no processing is performed.

[0147] Among them, the preset standard deviation needs to be set according to the actual detection situation, and no specific limitation is made here.

[0148] Here, the standard deviations of each category in the tracking image sequence are respectively compared with the preset standard deviation; when the standard deviation is less than the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is greater than the preset rotation angle, it is determined that the category corresponding to the standard deviation is a height undulation target signal, and the category is deleted from the tracking image sequence. It should be emphasized that when both determination conditions are adopted, only when both conditions that the standard deviation is less than the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is greater than the preset rotation angle are met can it be determined as a height undulation target signal, and no determination result can be obtained as long as one condition is not satisfied.

[0149] When the standard deviation is greater than or equal to the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is less than or equal to the preset rotation angle, it is determined that the category corresponding to the standard deviation is a moving target signal and no processing is required. Here, by differentiating the motion trajectories of the moving target signal and the height undulation target signal, the category corresponding to the height undulation target signal is deleted from the tracking image sequence, effectively reducing the false alarm rate of moving target detection and achieving the effect of improving the accuracy of moving target detection.

[0150] Combined with the implementation methods in the above embodiments, the following Figure 11 is an example description of a preferred method flow provided by the embodiments of the present application. As Figure 11 shown, the method may include the following steps:

[0151] Step S501, obtain circular synthetic aperture radar data, perform background difference processing on the circular synthetic aperture radar data to obtain an initial image sequence.

[0152] Step S502, perform morphological processing and clustering operation on the initial image sequence to obtain a clustered image sequence.

[0153] Step S503, obtain the category with the highest overlapping pixels in every two adjacent frames of images of the same target signal in the clustered image sequence.

[0154] Step S504, judge the number of categories with the highest overlapping pixels. When it is greater than 1, execute step S505; when it is equal to 1, execute step S509.

[0155] Step S505, then use the prefabricated cross ratio algorithm to calculate the pixel cross ratio in every two adjacent frames of images in each category with the highest overlapping pixels in the same target signal.

[0156] Step S506: Compare the calculated pixel cross ratios of each pixel to obtain the maximum pixel cross ratio.

[0157] Step S507: Compare the maximum pixel cross ratio with a preset cross ratio threshold. If the maximum pixel cross ratio is greater than the preset cross ratio threshold, change the category of the overlapping pixels in the subsequent frame image to the category of the overlapping pixels in the previous frame image.

[0158] Step S508: If the maximum pixel cross ratio is less than or equal to the preset cross ratio threshold, empty the category of the overlapping pixels in the subsequent frame image; execute Step S512.

[0159] Step S509: Then, use the prefabricated cross ratio algorithm to calculate the pixel cross ratios of every two adjacent frame images of the same target signal.

[0160] Step S510: Compare the pixel cross ratio with the preset cross ratio threshold. If the pixel cross ratio is greater than the preset cross ratio threshold, change the category of the overlapping pixels in the subsequent frame image to the category of the overlapping pixels in the previous frame image.

[0161] Step S511: If the pixel cross ratio is less than or equal to the preset cross ratio threshold, empty the category of the overlapping pixels in the subsequent frame image.

[0162] Step S512: And so on, to obtain a tracking image sequence with multiple categories corresponding one by one to the target signal.

[0163] Step S513: Based on the feature extraction function, extract the feature values of each frame image of each category in the tracking image sequence.

[0164] Step S514: Calculate the endpoint coordinate values of each frame image of each category according to the feature values of each category in the tracking image sequence.

[0165] Step S515: Calculate the feature angle of each frame image according to the endpoint coordinate values of each frame image of each category in the tracking image sequence.

[0166] Step S516: Subtract the feature angle of the first frame image of each category in the tracking image sequence from the feature angle of the last frame image to obtain the cumulative rotation angle of each category in the tracking image sequence; execute Step S519.

[0167] Step S517: According to the endpoint coordinate values of each frame image, obtain the center point coordinate values of each category in the tracking image sequence.

[0168] Step S518: According to the endpoint coordinate values of each frame image of each category and the corresponding center point coordinate values, obtain the standard deviation of each category in the tracking image sequence.

[0169] Step S519: Compare the cumulative rotation angles of each category in the tracking image sequence with a preset rotation angle respectively, and compare the standard deviations of each category in the tracking image sequence with a preset standard deviation respectively. When the cumulative rotation angle is greater than the preset rotation angle and the standard deviation is less than the preset standard deviation, execute Step S520. When the cumulative rotation angle is less than or equal to the preset rotation angle and the standard deviation is greater than or equal to the preset standard deviation, execute Step S521.

[0170] Step S520: Determine it as a height fluctuation target signal, and delete the category corresponding to the height fluctuation target signal from the tracking image sequence.

[0171] Step S521: Determine it as a moving target signal and do not perform any processing.

[0172] It should be understood that although Figure 2 、 Figure 5 、 Figure 6 and Figure 11 in the flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this application, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, Figure 2 、 Figure 5 、 Figure 6 and Figure 11 at least a part of the steps in can include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same moment, but can be executed at different moments. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0173] The above method embodiments can be applied to a variety of application scenarios. For example, they can include but are not limited to the following application scenarios:

[0174] Based on the image sequence collected by a circular synthetic aperture radar, distinguish the moving target signals and height fluctuation target signals in the image sequence. After determining the height fluctuation target signal, delete the category corresponding to the height fluctuation target signal from the image sequence to achieve the effect of improving the accuracy of moving target detection.

[0175] Figure 12 FIG. is a schematic structural diagram of a circular synthetic aperture radar moving target detection device provided by an embodiment of the present application. This device can be set in Figure 1 the server in the system shown in, and is used to execute the method processes shown in Figure 2 、 Figure 5 、 Figure 6 and Figure 11 in. AsFigure 12 As shown, the device may include: a background difference processing module 601, a morphology and clustering processing module 603, a target tracking module 605, a feature extraction module 607, and a target discrimination module 609. The main functions of each component module are as follows:

[0176] The background difference processing module 601 is used to obtain circumferential synthetic aperture radar data, perform background difference processing on the circumferential synthetic aperture radar data, and obtain an initial image sequence;

[0177] The morphology and clustering processing module 603 is used to perform morphological processing and clustering operations on the initial image sequence to obtain a clustering image sequence;

[0178] The target tracking module 605 is used to perform target tracking on the clustering image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals;

[0179] The feature extraction module 607 is used to extract the feature values of each category in the tracking image sequence, and calculate the cumulative rotation angles of each category in the tracking image sequence according to the feature values of each category;

[0180] The target discrimination module 609 is used to distinguish moving target signals and height fluctuation target signals according to the cumulative rotation angles of each category in the tracking image sequence, and delete the category corresponding to the height fluctuation target signal from the tracking image sequence.

[0181] In some embodiments, the target tracking module 605 further performs:

[0182] Obtain the category with the highest overlapping pixels in every two adjacent frames of images of the same target signal in the clustering image sequence;

[0183] If the number of categories with the highest overlapping pixels is equal to 1, use the prefabricated cross-ratio algorithm to calculate the pixel cross-ratio of every two adjacent frames of images;

[0184] Compare the pixel cross-ratio with a preset cross-ratio threshold. If the pixel cross-ratio is greater than the preset cross-ratio threshold, change the category of the overlapping pixels in the latter frame of the image to the category of the overlapping pixels in the former frame of the image;

[0185] If the pixel cross-ratio is less than or equal to the preset cross-ratio threshold, set the category of the overlapping pixels in the latter frame of the image to be empty, and so on, to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals.

[0186] In some embodiments, the target tracking module 605 specifically further performs:

[0187] Obtain the category with the highest overlapping pixels in every two adjacent frames of images of the same target signal in the clustering image sequence;

[0188] If the number of categories with the highest overlapping pixels is greater than 1, the prefabricated cross-ratio algorithm is used to calculate the pixel cross-ratio in each adjacent two frames of images in each category with the highest overlapping pixels in the same target signal respectively;

[0189] Compare the calculated pixel cross-ratios to obtain the maximum pixel cross-ratio;

[0190] Compare the maximum pixel cross-ratio with the preset cross-ratio threshold. If the maximum pixel cross-ratio is greater than the preset cross-ratio threshold, change the category of the overlapping pixels in the latter frame image to the category of the overlapping pixels in the former frame image;

[0191] If the maximum pixel cross-ratio is less than or equal to the preset cross-ratio threshold, empty the category of the overlapping pixels in the latter frame image, and so on, to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signal.

[0192] In some embodiments, the feature extraction module 607 also performs:

[0193] Extract the feature values of each frame of images in each category in the tracking image sequence based on the feature extraction function, and calculate the endpoint coordinate values of each frame of images in each category according to the feature values of each category in the tracking image sequence;

[0194] Calculate the feature angle of each frame of image according to the endpoint coordinate values of each frame of images in each category in the tracking image sequence;

[0195] Subtract the feature angle of the first frame of image in each category in the tracking image sequence from the feature angle of the last frame of image to obtain the cumulative rotation angle of each category in the tracking image sequence.

[0196] In some embodiments, the target discrimination module 609 specifically also performs:

[0197] Compare the cumulative rotation angles of each category in the tracking image sequence with the preset rotation angle respectively;

[0198] When the cumulative rotation angle is greater than the preset rotation angle, determine that the category corresponding to the cumulative rotation angle is a height fluctuation target signal, and delete the category from the tracking image sequence;

[0199] If the cumulative rotation angle is less than or equal to the preset rotation angle, determine that the category corresponding to the cumulative rotation angle is a moving target signal and do not perform any processing.

[0200] In some embodiments, the device also performs:

[0201] According to the endpoint coordinate values of each frame of image, obtain the center point coordinate values of each category in the tracking image sequence;

[0202] According to the endpoint coordinate values and corresponding center point coordinate values of each frame image of each category, the standard deviation of each category in the tracking image sequence is obtained;

[0203] According to the standard deviation of each category in the tracking image sequence, a moving target signal and a height fluctuation target signal are determined, and the category corresponding to the height fluctuation target signal is deleted from the tracking image sequence.

[0204] In some embodiments, the device further specifically performs:

[0205] The standard deviations of each category in the tracking image sequence are respectively compared with a preset standard deviation;

[0206] When the standard deviation is less than the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is greater than the preset rotation angle, it is determined that the category corresponding to the standard deviation is a height fluctuation target signal, and the category is deleted from the tracking image sequence;

[0207] When the standard deviation is greater than or equal to the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is less than or equal to the preset rotation angle, it is determined that the category corresponding to the standard deviation is a moving target signal, and no processing is performed.

[0208] For the same and similar parts among the above embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant parts, reference can be made to the description of the method embodiment.

[0209] It should be noted that the embodiments of the present application may involve the use of user data. In actual applications, user-specific personal data can be used in the solutions described herein within the scope permitted by applicable laws and regulations (for example, with the user's explicit consent, actual notice to the user, and explicit authorization of the user, etc.) in compliance with the requirements of applicable laws and regulations of the country where it is located.

[0210] According to the embodiments of the present application, the present application also provides a computer device and a computer-readable storage medium.

[0211] As Figure 13 shown, it is a block diagram of a computer device according to an embodiment of the present application. The computer device is intended to represent various forms of digital computers or mobile devices. Among them, the digital computer may include a desktop computer, a portable computer, a workbench, a personal digital assistant, a server, a mainframe computer, and other suitable computers. The mobile device may include a tablet computer, a smart phone, a wearable device, etc.

[0212] As Figure 13As shown, device 700 includes a computing unit 701, a ROM 702, a RAM 703, a bus 704, and an input / output (I / O) interface 705. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via bus Y04. The input / output (I / O) interface 705 is also connected to the bus 704.

[0213] The computing unit 701 can perform various processes in the method embodiments of this application according to computer instructions stored in the read-only memory (ROM) 702 or computer instructions loaded from the storage unit 708 into the random access memory (RAM) 703. The computing unit 701 can be various general and / or special processing components with processing and computing capabilities. The computing unit 701 can include, but is not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. In some embodiments, the method provided in the embodiments of this application can be implemented as a computer software program, which is tangibly contained in a computer-readable storage medium, such as the storage unit 708.

[0214] The RAM 703 can also store various programs and data required for the operation of the device 700. Part or all of the computer program can be loaded and / or installed onto the device 700 via the ROM 702 and / or the communication unit 709.

[0215] The input unit 706, the output unit 707, the storage unit 708, and the communication unit 709 in the device 700 can be connected to the I / O interface 705. Among them, the input unit 706 can be such as a keyboard, a mouse, a touch screen, a microphone, etc.; the output unit 707 can be such as a display, a speaker, an indicator light, etc. The device 700 can exchange information, data, etc. with other devices through the communication unit 709.

[0216] It should be noted that this device may also include other components necessary for normal operation. It may also only include the components necessary to implement the solution of this application, and does not necessarily include all the components shown in the figure.

[0217] The various embodiments of the systems and technologies described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof.

[0218] The computer instructions for implementing the method of the present application can be written in any combination of one or more programming languages. These computer instructions can be provided to the computing unit 701, so that when the computer instructions are executed by the computing unit 701 such as a processor, the steps involved in the method embodiments of the present application are executed.

[0219] The computer-readable storage medium provided by the present application can be a tangible medium, which can contain or store computer instructions for executing the steps involved in the method embodiments of the present application. The computer-readable storage medium can include, but is not limited to, storage media in the forms of electronic, magnetic, optical, electromagnetic, etc.

[0220] The above specific implementation manners do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A method for moving target detection of circular synthetic aperture radar, characterized in that, The method includes: Obtaining circumferential synthetic aperture radar data, performing background difference processing on the circumferential synthetic aperture radar data to obtain an initial image sequence; Performing morphological processing and clustering operation on the initial image sequence to obtain a clustering image sequence; Performing target tracking on the clustering image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to target signals; Extracting the feature values of each frame of image of each category in the tracking image sequence based on a feature extraction function, and calculating the endpoint coordinate values of each frame of image of each category according to the feature values of each category in the tracking image sequence; Obtaining the center point coordinate values of each category in the tracking image sequence according to the endpoint coordinate values of each frame of image; Obtaining the standard deviation of each category in the tracking image sequence according to the endpoint coordinate values of each frame of image of each category and the corresponding center point coordinate values; and, Calculating the feature angle of each frame of image according to the endpoint coordinate values of each frame of image of each category in the tracking image sequence; Calculating the difference between the feature angle of the first frame of image and the feature angle of the last frame of image of each category in the tracking image sequence to obtain the cumulative rotation angle of each category in the tracking image sequence; Distinguishing moving target signals and height fluctuation target signals according to the standard deviation and the corresponding cumulative rotation angle of each category in the tracking image sequence, and deleting the category corresponding to the height fluctuation target signal from the tracking image sequence.

2. The method according to claim 1, wherein The performing target tracking on the clustering image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to target signals includes: Obtaining the category with the highest overlapping pixels in every two adjacent frames of images of the same target signal in the clustering image sequence; If the number of categories with the highest overlapping pixels is equal to 1, calculating the pixel cross-ratio of every two adjacent frames of images by using a prefabricated cross-ratio algorithm; Comparing the pixel cross-ratio with a preset cross-ratio threshold, if the pixel cross-ratio is greater than the preset cross-ratio threshold, changing the category of the overlapping pixels in the latter frame of image to the category of the overlapping pixels in the former frame of image; If the pixel cross-ratio is less than or equal to the preset cross-ratio threshold, emptying the category of the overlapping pixels in the latter frame of image, and so on, to obtain a tracking image sequence with multiple categories corresponding one-to-one to target signals.

3. The method according to claim 1, wherein The performing target tracking on the clustering image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to target signals includes: Obtaining the category with the highest overlapping pixels in every two adjacent frames of images of the same target signal in the clustering image sequence; If the number of categories with the highest overlapping pixels is greater than 1, calculating the pixel cross-ratios of every two adjacent frames of images in each category with the highest overlapping pixels in the same target signal respectively by using a prefabricated cross-ratio algorithm; Comparing the calculated pixel cross-ratios to obtain the maximum pixel cross-ratio; Comparing the maximum pixel cross-ratio with a preset cross-ratio threshold, if the maximum pixel cross-ratio is greater than the preset cross-ratio threshold, changing the category of the overlapping pixels in the latter frame of image to the category of the overlapping pixels in the former frame of image; If the maximum pixel cross-ratio is less than or equal to the preset cross-ratio threshold, the categories of overlapping pixels in the subsequent frame image are emptied, and so on, to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals.

4. The method according to any one of claims 1 to 3, characterized in that The distinguishing of moving target signals and height-undulating target signals according to the standard deviations and corresponding cumulative rotation angles of each category in the tracking image sequence, and deleting the category corresponding to the height-undulating target signal from the tracking image sequence includes: comparing the standard deviations of each category in the tracking image sequence with a preset standard deviation respectively, and comparing the cumulative rotation angles of each category in the tracking image sequence with a preset rotation angle respectively; When the standard deviation is less than the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is greater than the preset rotation angle, it is determined that the category corresponding to the standard deviation is a height-undulating target signal, and the category is deleted from the tracking image sequence; When the standard deviation is greater than or equal to the preset standard deviation and the cumulative rotation angle of the category corresponding to the standard deviation is less than or equal to the preset rotation angle, it is determined that the category corresponding to the standard deviation is a moving target signal and no processing is performed.

5. A moving target detection device for circular synthetic aperture radar, characterized in that, The device includes: a background difference processing module, configured to obtain circumferential synthetic aperture radar data, perform background difference processing on the circumferential synthetic aperture radar data to obtain an initial image sequence; a morphological and clustering processing module, configured to perform morphological processing and clustering operation on the initial image sequence to obtain a clustered image sequence; a target tracking module, configured to perform target tracking on the clustered image sequence to obtain a tracking image sequence with multiple categories corresponding one-to-one to the target signals; a feature extraction module, configured to extract the feature values of each frame image of each category in the tracking image sequence based on a feature extraction function, calculate the endpoint coordinate values of each frame image of each category according to the feature values of each category in the tracking image sequence; obtain the center point coordinate values of each category in the tracking image sequence according to the endpoint coordinate values of each frame image; obtain the standard deviations of each category in the tracking image sequence according to the endpoint coordinate values of each frame image of each category and the corresponding center point coordinate values; and calculate the feature angle of each frame image according to the endpoint coordinate values of each frame image of each category in the tracking image sequence; calculate the difference between the feature angle of the first frame image and the feature angle of the last frame image of each category in the tracking image sequence to obtain the cumulative rotation angle of each category in the tracking image sequence; a target discrimination module, configured to distinguish moving target signals and height-undulating target signals according to the standard deviations and corresponding cumulative rotation angles of each category in the tracking image sequence, and delete the category corresponding to the height-undulating target signal from the tracking image sequence.

6. A computer device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores computer instructions executable by the at least one processor, and the computer instructions are executed by the at least one processor to enable the at least one processor to execute the method according to any one of claims 1-4.

7. A computer-readable storage medium having computer instructions stored thereon, characterized in that, The computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 4.

Images