VTS radar target extraction method based on video analysis

Abstract

The invention relates to the technical field of data processing, and discloses a VTS radar target extraction method based on video analysis, which comprises the following steps of: acquiring an original echo image of an ith frame; and preprocessing the echo image of the ith frame. And obtaining the preprocessed echo image of the ith frame, and carrying out binarization processing on the image. And performing closed operation processing on the image after binarization processing, and removing part of noisy points and holes in the image. And performing target contour extraction on the image after closed operation processing to obtain contour information of each target. And obtaining the original echo image of the next frame, obtaining the target contour information of the (i + 1) frame according to the processing method of the above steps, calculating the similarity between the target contour information and the target contour of the target contour library, and determining the target contour of the (i + 1) frame according to the similarity. According to the technical scheme provided by the invention, the precision of target extraction can be improved, the risk of target splitting is reduced, and a new thought is provided for VTS radar target extraction.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a VTS radar target extraction method based on video analysis. Background Art

[0002] With the rapid development of my country's coastal areas, the number of ships at sea is increasing, and the traffic management of ships at sea has become particularly important. In order to better manage the ships coming and going, more and more ports and waters choose to install VTS radar systems.

[0003] However, radar detection of targets is sometimes discontinuous, resulting in one target being split into multiple targets, reducing the integrity of target extraction, thereby affecting the accuracy of target point condensation and causing unstable target tracking. Summary of the invention

[0004] The technical problem solved by the present invention is that radar detection of targets is sometimes discontinuous, resulting in one target being split into multiple targets, thereby reducing the integrity of target extraction.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] A VTS radar target extraction method based on video analysis comprises the following steps:

[0007] S1. Obtain the original echo image of the i-th frame, where i represents the frame number scanned by the VTS radar, i=1, 2, 3...n.

[0008] S2. Preprocess the original echo image of the i-th frame.

[0009] S3. Obtain the echo image of the i-th frame after preprocessing, and perform binarization processing on the image.

[0010] S4. Perform a closing operation on the binarized image to remove some noise points and holes in the image.

[0011] S5. Extract target contours from the image processed by the closing operation to obtain contour information of each target.

[0012] S6. Obtain the original echo image of the next frame, and obtain the target contour information of the i+1 frame according to the processing method of steps S2-S5, and calculate the similarity with the target contour in the target contour library, and determine the target contour of the i+1 frame based on the similarity.

[0013] The VTS radar target extraction method based on video analysis according to the present invention has the following technical effects: the accuracy of target extraction can be improved, the risk of target splitting can be reduced, and a new idea for VTS radar target extraction can be provided.

[0014] In one embodiment of the present invention, in step S1, the process of obtaining the original echo image includes: taking the azimuth information as the vertical axis and the distance library information as the horizontal axis, and using the amplitude value of each distance library at each azimuth as the pixel value of the pixel point corresponding to the distance library at that azimuth to form a two-dimensional grayscale image.

[0015] In one embodiment of the present invention, in step S2, the preprocessing process includes: screening the pixel points whose pixel values ​​in the original echo image are lower than the preset clutter threshold and setting them to 0, so as to obtain the preprocessed image G f .

[0016] In one embodiment of the present invention, in step S3, the binarization process includes: evenly dividing the distance library information into three segments, which are recorded as x1, x2, and x3 from near to far. Setting a threshold for each segment, which are recorded as T1, T2, and T3 respectively. Determining whether the pixel value of each pixel is greater than or equal to the corresponding threshold, if yes, the pixel value of the pixel in the preprocessed image is set to 255, otherwise, the pixel value of the pixel is set to 0.

[0017] In one embodiment of the present invention, in step S4, the closing operation process includes: selecting a 3×3 rectangle as the structural element, and setting the domain of the structural element to D S . Set the domain of the binary image to D G The binary image is subjected to dilation, erosion and closing operations in sequence.

[0018] In one embodiment of the present invention, the formula for the expansion operation is as follows:

[0019]

[0020] In the formula, G thres is the image after binarization, and S(x,y) is the structural element.

[0021] In one embodiment of the present invention, the formula for the corrosion operation is as follows:

[0022] G thres ⊙S(i,j)=min{G thres (i+x,j+y)-S(x,y)|(i+x,j+y)∈D G , (x,y)∈D S}.

[0023] In one embodiment of the present invention, the formula for the closing operation is as follows:

[0024]

[0025] In one embodiment of the present invention, step S6 includes the following specific steps:

[0026] S61, according to the contour information of the same target in the i-2th frame, the i-1th frame, and the i-th frame in terms of distance and direction, respectively obtain the minimum circumscribed rectangle of the contour, sort according to the sum of the length and width of the rectangle, remove the lowest and highest values, and select the target contour of the frame corresponding to the middle value as A.

[0027] S62, count the Hu moment features of contour A, and use the Hu moment features, the width and height of the minimum circumscribed rectangle of the contour as the feature library of the target. Through steps S2-S5, the contour information of the target in the i+1 frame is obtained, recorded as B, and its Hu moment features, the width and height of the minimum circumscribed rectangle of the contour are counted. If the sum of the width and height of the minimum circumscribed rectangle of contour B is greater than the corresponding value in the feature library of the target, contour B is directly used as the contour information of the i+1 frame of the target. Otherwise, the Hu moment features in the feature library of the target are compared to calculate the similarity SIM of the two contours;

[0028] S63: If SIM is smaller than a preset threshold, it means that the target contour is matched successfully, and the contour B is used as the contour information of the target in the i+1th frame.

[0029] S64, if SIM is greater than a preset threshold, it means that the target contour matching fails. According to the minimum bounding rectangle of the target contour in the i+1 frame, in the image obtained in step S2, the enlarged sub-image of the target is intercepted. According to the width and height of the minimum bounding rectangle of the contour of the target feature library, the sub-image is subjected to cyclic multi-scale threshold segmentation, and the edge contour of the sub-image is extracted. The binarization threshold is gradually reduced until the sum of the width and height of the contour reaches the sum of the width and height of the target feature library, and the contour at this time is used as the contour information of the target i+1 frame.

[0030] In one embodiment of the present invention, the calculation formula of the similarity SIM is as follows:

[0031]

[0032] In the formula, and are the i-th moment of the Hu moment of contour A and contour B respectively, i = 1, 2,…, 7.

[0033] Beneficial effects of the present invention:

[0034] (1) The segmented binarization processing of the present invention effectively filters out background clutter and improves the target extraction accuracy; the closed operation processing of the present invention can remove some noise points and holes in the target and improve the integrity of target extraction;

[0035] (2) The present invention combines the target information of the first three frames of history, and effectively improves the target extraction capability by performing a series of processes such as contour matching judgment, contour size judgment, and multi-scale threshold segmentation on the target;

[0036] (3) The present invention reduces the risk of the same target being split into multiple targets during target extraction through a series of algorithms, thus providing a reliable basis for subsequent point track processing and track processing;

[0037] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0039] Figure 1 It is a flow chart of a VTS radar target extraction method based on video analysis of the present invention;

[0040] Figure 2 It is a flowchart of the VTS radar target extraction method based on video analysis of the present invention;

[0041] Figure 3 is the original target echo image of the i-th frame of the present invention;

[0042] Figure 4 The result image of the i-th frame original target echo image after preprocessing of the present invention;

[0043] Figure 5 The result graph of binarization of the i-th frame preprocessed image of the present invention;

[0044] Figure 6 A result image of a closing operation performed on the i-th frame of the binary image of the present invention;

[0045] Figure 7 This is a result diagram of target contour extraction of the image after the closing operation of the i-th frame of the present invention.

[0046] Figure 8 This is a result diagram of target contour extraction of the original target echo image of the i+1th frame of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In the description of the present invention, it should be understood that descriptions involving orientations, such as up, down, front, back, left, right, etc., and orientations or positional relationships indicated are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0049] In the description of the present invention, "several" means one or more, "more" means more than two, "greater than", "less than", "exceed" etc. are understood as not including the number itself, and "above", "below", "within" etc. are understood as including the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing the technical features, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0050] See also Figure 1-8 The present invention is a VTS radar target extraction method based on video analysis, comprising the following steps:

[0051] S1. Obtain the original echo image of the i-th frame, where i represents the frame number scanned by the VTS radar, i=1, 2, 3...n.

[0052] S2. Preprocess the original echo image of the i-th frame.

[0053] S3. Obtain the echo image of the i-th frame after preprocessing, and perform binarization processing on the image.

[0054] S4. Perform a closing operation on the binarized image to remove some noise points and holes in the image.

[0055] S5. Extract target contours from the image processed by the closing operation to obtain contour information of each target.

[0056] S6. Obtain the original echo image of the next frame, and obtain the target contour information of the i+1 frame according to the processing method of steps S2-S5, and calculate the similarity with the target contour in the target contour library, and determine the target contour of the i+1 frame based on the similarity.

[0057] It should be supplemented that, in some embodiments of the present invention, during one frame of VTS radar scanning, the radar monitoring azimuth ranges from 0° to 360°, with due north as 0°, the azimuth sampling interval is 0.08789°, and the detected azimuth is quantified into 4096 azimuths; the radar detection distance ranges from 0km to 20km, the detection distance is relative to the radar position, the distance sampling interval is 9.77m, and the detection distance is quantified into 4096 distance libraries; that is, the radar starts from the 0° azimuth, receives the target echo information on the 4096 distance libraries on the 0° azimuth, and then receives the target echo information on the 4096 distance libraries on the next azimuth in sequence according to the azimuth sampling interval of 0.08789°.

[0058] It should be noted that in step S1, the process of obtaining the original echo image includes: taking the azimuth information as the vertical axis and the distance library information as the horizontal axis, the amplitude value of each distance library at each azimuth is used as the pixel value of the pixel point corresponding to the distance library at the azimuth to form a two-dimensional grayscale image. For example, considering the real-time processing problem, the data of each 90° azimuth is accumulated to form an image, that is, the 360° azimuth is divided into 4 sub-images for separate processing, and the size of each image is 1024*2048.

[0059] It should be noted that in step S2, the preprocessing process includes: screening the pixel points in the original echo image whose pixel values ​​are lower than the preset clutter threshold and setting them to 0 to obtain the preprocessed image G f .

[0060] It should be noted that in step S3, the preprocessed image G f Binarization is performed to obtain a binary image G thres The binarization process includes: evenly dividing the distance library information into 3 segments, recorded as x1, x2, and x3 from near to far. Setting a threshold for each segment, recorded as T1, T2, and T3 respectively. Determine whether the pixel value of each pixel is greater than or equal to the corresponding threshold. If so, set the pixel value of the pixel in the preprocessed image to 255, otherwise set the pixel value of the pixel to 0.

[0061]

[0062] It should be noted that in step S4, the closing operation process includes: selecting a 3×3 rectangle as the structural element and setting the domain of the structural element to D S . Set the domain of the binary image to D G The binary image G of the i-th frame is subjected to dilation, erosion and closing operations in sequence. thres Perform a closing operation to obtain image G close; Dilation is to calculate the maximum value of the pixel points in the area covered by the structural element, and assign the maximum value to the pixel specified by the reference point; Erosion is to calculate the minimum value of the pixel points in the area covered by the structural element, and assign the minimum value to the pixel specified by the reference point.

[0063] The formula for the expansion operation is as follows:

[0064]

[0065] In the formula, G thres is the image after binarization, and S(x,y) is the structural element.

[0066] The formula for the corrosion operation is as follows:

[0067] G thres ⊙S(i,j)=min{G thres (i+x,j+y)-S(x,y)|(i+x,j+y)∈D G , (x,y)∈D S}.

[0068] The formula for the closing operation is as follows:

[0069]

[0070] It should be noted that in step S6, the contour information of each target in the next frame is obtained by combining the historical frame image. Taking one target as an example, the same is true for other targets. The specific method is as follows:

[0071] S61, according to the contour information of the same target in the i-2th frame, the i-1th frame, and the i-th frame in terms of distance and direction, respectively obtain the minimum circumscribed rectangle of the contour, sort according to the sum of the length and width of the rectangle, remove the lowest and highest values, and select the target contour of the frame corresponding to the middle value as A.

[0072] S62, count the Hu moment features of contour A, and use the Hu moment features, the width and height of the minimum circumscribed rectangle of the contour as the feature library of the target. Through steps S2-S5, the contour information of the target in the i+1 frame is obtained, recorded as B, and its Hu moment features, the width and height of the minimum circumscribed rectangle of the contour are counted. If the sum of the width and height of the minimum circumscribed rectangle of contour B is greater than the corresponding value in the feature library of the target, contour B is directly used as the contour information of the i+1 frame of the target. Otherwise, the Hu moment features in the feature library of the target are compared to calculate the similarity SIM of the two contours;

[0073] S63: If SIM is smaller than a preset threshold, it means that the target contour is matched successfully, and the contour B is used as the contour information of the target in the i+1th frame.

[0074] S64, if SIM is greater than a preset threshold, it means that the target contour matching fails. According to the minimum bounding rectangle of the target contour in the i+1 frame, in the image obtained in step S2, the enlarged sub-image of the target is intercepted. According to the width and height of the minimum bounding rectangle of the contour of the target feature library, the sub-image is subjected to cyclic multi-scale threshold segmentation, and the edge contour of the sub-image is extracted. The binarization threshold is gradually reduced until the sum of the width and height of the contour reaches the sum of the width and height of the target feature library, and the contour at this time is used as the contour information of the target i+1 frame.

[0075] The calculation formula of similarity SIM is as follows:

[0076]

[0077] In the formula, and are the i-th moment of the Hu moment of contour A and contour B respectively, i = 1, 2,…, 7.

[0078] The advantages of the present invention are:

[0079] (1) The segmented binarization processing of the present invention effectively filters out background clutter and improves the target extraction accuracy; the closed operation processing of the present invention can remove some noise points and holes in the target and improve the integrity of target extraction;

[0080] (2) The present invention combines the target information of the first three frames of history, and effectively improves the target extraction capability by performing a series of processes such as contour matching judgment, contour size judgment, and multi-scale threshold segmentation on the target;

[0081] (3) The present invention reduces the risk of the same target being split into multiple targets during target extraction through a series of algorithms, thus providing a reliable basis for subsequent point processing and track processing.

[0082] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of application of the present invention should still fall within the scope of the claims of the present invention.

Claims

1. A VTS radar target extraction method based on video analysis, characterized in that: The following steps are involved: S1, obtaining the original echo image of the i-th frame; Where i represents the number of frames scanned by the VTS radar, i = 1, 2, 3…n; S2, preprocessing the original echo image of the i-th frame; S3, obtaining the echo image of the i-th frame after preprocessing, and performing binarization processing on the image; S4, performing a closing operation on the image after the binary processing to remove some noise points and holes in the image; S5, extracting the target contour from the image processed by the closing operation to obtain the contour information of each target; S6. Obtain the original echo image of the next frame, and obtain the target contour information of the i+1 frame according to the processing method of steps S2-S5, and calculate the similarity with the target contour in the target contour library, and determine the target contour of the i+1 frame based on the similarity.

2. The VTS radar target extraction method based on video analysis according to claim 1, characterized in that: In step S1, the process of obtaining the original echo image includes: taking the azimuth information as the vertical axis and the distance library information as the horizontal axis, and using the amplitude value of each distance library at each azimuth as the pixel value of the pixel point corresponding to the distance library at the azimuth to form a two-dimensional grayscale image.

3. The VTS radar target extraction method based on video analysis according to claim 1, characterized in that: In step S2, the preprocessing process includes: screening the pixel points in the original echo image whose pixel values ​​are lower than the preset clutter threshold and setting them to 0, so as to obtain the preprocessed image G f .

4. The VTS radar target extraction method based on video analysis according to claim 1, characterized in that: In step S3, the binarization process includes: evenly dividing the distance library information into 3 segments, recorded as x1, x2, and x3 from near to far; setting a threshold for each segment, recorded as T1, T2, and T3 respectively; judging whether the pixel value of each pixel point is greater than or equal to the corresponding threshold, if so, setting the pixel value of the pixel point in the preprocessed image to 255, otherwise, setting the pixel value of the pixel point to 0.

5. The VTS radar target extraction method based on video analysis according to claim 1, characterized in that: In step S4, the closing operation process includes: selecting a 3×3 rectangle as the structural element, and setting the domain of the structural element to D S ; Set the domain of the binary image to D G ; Perform dilation, erosion and closing operations on the binary image in sequence.

6. The VTS radar target extraction method based on video analysis according to claim 5 is characterized in that: The formula for the expansion operation is as follows: In the formula, G thres is the image after binarization, and S(x,y) is the structural element.

7. The VTS radar target extraction method based on video analysis according to claim 6 is characterized in that: The formula for the corrosion operation is as follows: G thres ⊙S(i,j)=min{G thres (i+x,j+y)-S(x,y)|(i+x,j+y)∈D G ,(x,y)∈D S }。 8. The VTS radar target extraction method based on video analysis according to claim 7 is characterized in that: The formula for the closing operation is as follows:

9. The VTS radar target extraction method based on video analysis according to claim 1, characterized in that: Step S6 includes the following specific steps: S61, according to the contour information of the same target in the distance and azimuth of the i-2th frame, the i-1th frame, and the i-th frame, respectively obtain the minimum circumscribed rectangle of the contour; sort according to the sum of the length and width of the rectangle, remove the lowest and highest values, and select the target contour of the frame corresponding to the middle value and record it as A; S62, counting the Hu moment features of contour A, and using the Hu moment features, the width and height of the minimum circumscribed rectangle of the contour as the feature library of the target; obtaining the contour information of the target in the i+1th frame through steps S2-S5, recorded as B, and counting its Hu moment features, the width and height of the minimum circumscribed rectangle of the contour; if the sum of the width and height of the minimum circumscribed rectangle of contour B is greater than the corresponding value of the target feature library, then directly using contour B as the contour information of the i+1th frame of the target; otherwise, comparing the Hu moment features in the target feature library, and calculating the similarity SIM of the two contours; S63, if SIM is less than a preset threshold, it means that the target contour is matched successfully, and the contour B is used as the contour information of the target in the i+1th frame; S64. If SIM is greater than a preset threshold, it indicates that the target contour matching fails; based on the minimum bounding rectangle of the target contour in the i+1th frame, in the image obtained in step S2, an enlarged sub-image of the target is captured; based on the width and height of the minimum bounding rectangle of the contour in the target feature library, the sub-image is subjected to cyclic multi-scale threshold segmentation, and the edge contour of the sub-image is extracted; the binarization threshold is gradually reduced until the sum of the contour width and height reaches the sum of the width and height of the target feature library, and the contour at this time is used as the contour information of the target in the i+1th frame.

10. The VTS radar target extraction method based on video analysis according to claim 9 is characterized in that: The calculation formula of similarity SIM is as follows: In the formula, and are the i-th moment of the Hu moment of contour A and contour B respectively, i = 1, 2,…, 7.

Images