A cylindrical panoramic splicing method of steel wire rope

Abstract

The application discloses a cylindrical panoramic splicing method of a steel wire rope, and the method comprises the following steps: keeping the distance between a camera and the steel wire rope unchanged, rotating the camera around the steel wire rope at a constant speed for one round, and shooting the surface of the steel wire rope from directly above to obtain a steel wire rope video; then selecting effective frames of the steel wire rope video and extracting a steel wire rope strip-shaped graph according to the effective frames; finally obtaining a splicing area of the steel wire rope strip-shaped graph on a panoramic graph, and adding the splicing area of the steel wire rope strip-shaped graph to the panoramic graph, so that the cylindrical panoramic splicing of the steel wire rope is realized.

Description

Technical Field

[0001] This invention belongs to the field of non-destructive testing technology for steel wire ropes, and more specifically, it relates to a method for panoramic splicing of cylindrical steel wire ropes. Background Technology

[0002] Steel wire ropes are widely used in industrial settings, frequently serving as load-bearing, lifting, and structural anchoring ropes. The structural integrity of the steel wire rope significantly impacts the stability and reliability of the entire system. Long-term use often results in structural damage such as corrosion, wear, wire skipping, and broken wires, potentially leading to rope breakage and causing substantial property damage and loss of life. Therefore, regular inspection of steel wire ropes is essential to prevent industrial accidents.

[0003] Traditional manual visual inspection methods for steel wire ropes suffer from high labor costs and are significantly affected by subjective factors. To better detect defects in steel wire ropes, methods such as magnetic flux leakage detection, image recognition detection, and ultrasonic testing have been proposed. Among these, magnetic flux leakage detection is widely used in industrial settings due to its high reliability and recognition rate. However, magnetic flux leakage detection has drawbacks, including its inapplicability to steel wire ropes of varying thicknesses and its susceptibility to pull-out values.

[0004] In recent years, with the gradual improvement of computer performance, image detection methods for steel wire rope defects based on computer vision have received increasing attention. Computer vision methods typically involve image processing and feature extraction of acquired steel wire rope images, followed by classification using machine learning algorithms. However, existing methods typically capture images of the steel wire rope from the front, and the cylindricity of the steel wire rope causes cylindrical distortion at the edges of the cylinder in the resulting images. Furthermore, the captured images only provide partial information about the cylindrical surface. While these issues can be mitigated by capturing multiple images of the steel wire rope around its circumference, this increases the detection time cost of the recognition algorithm. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for panoramic stitching of the cylindrical surface of a steel wire rope. This method involves extracting and stitching features from a video of the steel wire rope taken around its length to obtain a panoramic information map of the steel wire rope's cylindrical surface.

[0006] To achieve the above-mentioned objectives, the present invention provides a method for panoramic splicing of steel wire ropes on a cylindrical surface, characterized by comprising the following steps:

[0007] (1) Filming a video of the steel wire rope;

[0008] (1.1) Set up two white LED light sources. The two white LED light sources are fixed in front of and behind the camera respectively along the axis of the steel wire rope. By adjusting the intensity of the two white LED light sources and the focal length of the camera, the surface texture of the steel wire rope can be displayed in the camera. Then record the distance between the camera and the steel wire rope in this state.

[0009] (1.2) Keep the distance between the camera and the wire rope constant, rotate the camera around the wire rope at a constant speed for one revolution, and shoot the surface of the wire rope from directly above to obtain the video of the wire rope;

[0010] (2) Select the effective frames of the steel wire rope video;

[0011] (2.1) Binarize each frame of the wire rope video, then use the Canny algorithm to perform edge detection on the binarized image to extract edge pixels. Finally, count the number of edge pixels in each frame, denoted as m. i , i = 1, 2, ..., N, where N represents the total number of frames in the wire rope video;

[0012] (2.2) Divide each frame of the wire rope video into several groups of image frames in chronological order. Each group includes n frames. If the total number of frames N is not a multiple of n, the number of frames in the last group shall be based on the remaining number of frames.

[0013] (2.3) For each group of image frames, take the maximum number of edge pixels (max(m)). k One frame of the image is taken as the valid frame in this group of image frames, k = 1, 2, ..., n;

[0014] (3) Obtain the wire rope strip diagram;

[0015] (3.1) For each group of image frames, extract the image after edge detection by the Canny algorithm corresponding to each valid frame. Treat each frame image as a matrix, and then count the number of edge pixels in each column of each matrix to obtain a row of statistical vectors. Let T be the statistical vector corresponding to the valid frame in the j-th group. j ,

[0016] (3.2) The statistical vector T j Each element in the matrix is ​​compared with a set threshold T'. The sequence number of the first element greater than or equal to T' to the last element greater than or equal to T' in the corresponding matrix is ​​found. Then, the corresponding column position in the corresponding valid frame is found according to the sequence number and the image is segmented. Finally, the image part with the corresponding sequence number is extracted as the pure steel wire rope image of the corresponding valid frame.

[0017] (3.3) Calculate the average width d of all pure steel wire images, and then cut a steel wire rope strip with a width of d / 3 from the center of each pure steel wire rope image;

[0018] (4) Obtain the stitching area contributed to the panoramic image by the steel wire rope strip image;

[0019] (4.1) Calculate the maximum number of columns to be joined, Δx max ;

[0020]

[0021] (4.2) Starting from the second wire rope strip image, calculate the similarity index between the current wire rope strip image and the previous wire rope strip image; then find the overlapping area with the highest similarity index between the current wire rope strip image and the previous wire rope strip image, and use it as the common part of the two wire rope strip images, while the remaining part of the current wire rope strip image is used as the stitching area contributed to the panoramic image;

[0022] (5) Steel wire rope strip splicing;

[0023] (5.1) Construct two grayscale mask matrices with the same number of rows as the wire rope strip diagram and k columns, where, The first grayscale mask matrix is ​​used on the current wire rope strip image, and the grayscale values ​​of each row are k numbers that gradually increase from 0 to 1. The second grayscale mask matrix is ​​used on the wire rope strip image of the previous frame, and the grayscale values ​​of each row are k numbers that gradually decrease from 1 to 0.

[0024] (5.2) Use the entire first wire rope strip diagram as the splicing area to form the starting panoramic image to be spliced;

[0025] (5.3) Starting from the second wire rope strip image, take the last k columns of the non-stitched area of ​​the current wire rope strip image and the first k columns of the current panoramic image as the stitching transition area, multiply them by the first and second grayscale mask matrices respectively, and then add them together. The result of the sum replaces the first k columns of the current panoramic image. Finally, add the stitching area of ​​the current wire rope strip image to the beginning of the current panoramic image.

[0026] (5.4) Repeat step (5.3) until all the steel wire rope strip diagrams are used up, thereby achieving the panoramic splicing of the steel wire rope cylindrical surface.

[0027] The objective of this invention is achieved as follows:

[0028] This invention discloses a method for panoramic stitching of a steel wire rope on a cylindrical surface. While maintaining a constant distance between the camera and the steel wire rope, the camera rotates uniformly around the steel wire rope for one revolution and captures a video of the steel wire rope surface from directly above. Then, valid frames from the steel wire rope video are selected, and a strip image of the steel wire rope is extracted from these frames. Finally, the stitching area contributed by the steel wire rope strip image to the panoramic image is obtained, and this stitching area is added to the panoramic image, thereby achieving panoramic stitching of the steel wire rope on a cylindrical surface.

[0029] Meanwhile, the cylindrical panoramic splicing method of steel wire rope of the present invention also has the following beneficial effects:

[0030] (1) This invention provides a novel method for detecting defects in steel wire rope images, which avoids cylindrical distortion in some areas caused by traditional steel wire rope images;

[0031] (2) The present invention extracts effective frames from the wire rope video to obtain pure wire rope images, and pure wire rope images have richer reflected light, thereby enhancing the image difference between the surface defects and normal parts of the wire rope.

[0032] (3) The present invention stitches together the information of the steel wire rope around the whole circle into a panoramic image, thereby avoiding the impact of recognizing multiple images taken from different angles on the time cost of the recognition algorithm. Attached Figure Description

[0033] Figure 1 This is a flowchart of a method for panoramic splicing of steel wire rope on a cylindrical surface according to the present invention;

[0034] Figure 2 This is a schematic diagram of the LED light source setup;

[0035] Figure 3 This is a diagram illustrating the video recording of the steel wire rope.

[0036] Figure 4 This is a schematic diagram illustrating the effective frame extraction from the wire rope video.

[0037] Figure 5 This is a schematic diagram of the steel wire rope strips;

[0038] Figure 6 This is a schematic diagram showing all the wire rope strip diagrams arranged in chronological order;

[0039] Figure 7 This is a diagram illustrating the maximum number of columns that can be joined.

[0040] Figure 8 It is a panoramic image created by stitching together the cylindrical surfaces of the steel wire rope. Detailed Implementation

[0041] The specific embodiments of the present invention will now be described with reference to the accompanying drawings to enable those skilled in the art to better understand the invention. It should be particularly noted that in the following description, detailed descriptions of known functions and designs that might obscure the main content of the invention will be omitted here.

[0042] Example

[0043] Figure 1 This is a flowchart of a method for panoramic splicing of steel wire rope on a cylindrical surface according to the present invention.

[0044] In this embodiment, as Figure 1 As shown, the present invention provides a method for panoramic splicing of steel wire rope on a cylindrical surface, comprising the following steps:

[0045] S1. Filming a video of the steel wire rope;

[0046] S1.1 Two white LED light sources are installed, fixed in front of and behind the camera respectively along the axis of the steel wire rope. Figure 2 As shown, by adjusting the intensity of the two white LED light sources and the camera focal length, the surface texture of the steel wire rope can be displayed in the camera, and then the distance between the camera and the steel wire rope in this state is recorded.

[0047] S1.2. Maintain a constant distance between the camera and the steel cable, such as... Figure 3 As shown, the camera rotates at a constant speed around the steel wire rope for one revolution and captures a video of the steel wire rope surface from directly above. In this embodiment, to avoid motion blur in the stitched footage, the camera's exposure time needs to be adjusted to a low value.

[0048] S2. Select valid frames from the wire rope video;

[0049] S2.1. Binarize each frame of the wire rope video, then use the Canny algorithm to perform edge detection on the binarized image to extract edge pixels. Finally, count the number of edge pixels in each frame, denoted as m. i i = 1, 2, ..., N, where N represents the total number of frames in the wire rope video. In this embodiment, the total number of frames is 890.

[0050] S2.2 Divide each frame of the wire rope video into several groups of image frames in chronological order. Each group includes n frames. If the total number of frames N is not a multiple of n, the number of frames in the last group is based on the remaining number of frames. In this embodiment, each frame of the video is divided into groups of 5 images in chronological order. If the total number of frames is not a multiple of 5, the number of frames in the last group may not be 5.

[0051] S2.3, such as Figure 4As shown, for each group of image frames, the maximum number of edge pixels (max(m)) is taken. k One frame of the image is taken as the valid frame in this group of image frames, k = 1, 2, ..., 5;

[0052] S3. Obtain the wire rope strip diagram;

[0053] S3.1 For each group of image frames, extract the image corresponding to each valid frame after edge detection using the Canny algorithm. Treat each frame as a matrix, and then count the number of edge pixels in each column of each matrix to obtain a row of statistical vectors. Let T be the statistical vector corresponding to the valid frame in the j-th group. j ,

[0054] S3.2, Calculate the statistical vector T j Each element in the matrix is ​​compared with the set threshold T' = 20. The sequence number of the first element greater than or equal to T' to the last element greater than or equal to 20 in the corresponding matrix is ​​found. Then, the corresponding column position in the corresponding valid frame is found according to the sequence number and the image is segmented. Finally, the image part with the corresponding sequence number is extracted as the pure steel wire rope image of the corresponding valid frame.

[0055] S3.3 Calculate the average width d of all pure steel wire images, such as Figure 5 As shown, a strip of wire rope with a width of d / 3 is cropped from the center of each pure wire rope image, as follows. Figure 6 As shown, the obtained wire rope strip diagrams are arranged in chronological order according to the effective frames, resulting in... Figure 7 .

[0056] S4. Obtain the stitching area on the panoramic image from the steel wire rope strip image;

[0057] S4.1 Calculate the maximum number of columns to be joined, Δx max ;

[0058]

[0059] In this embodiment, the maximum number of spliced ​​columns is 20;

[0060] S4.2, such as Figure 7 As shown, starting from the second wire rope bar chart, the similarity index between the current wire rope bar chart and the previous wire rope bar chart is calculated. Then, the overlapping area with the highest similarity index between the current wire rope bar chart and the previous wire rope bar chart is found and used as the common part of the two wire rope bar charts. The remaining part of the current wire rope bar chart is used as the stitching area contributed to the panoramic image. In this embodiment, the number of rows in the contributing part of the current wire rope bar chart is the same as that of the bar chart, and the number of columns does not exceed the number of updated stitching columns.

[0061] S5, Steel wire rope strip diagram splicing;

[0062] S5.1 Construct two grayscale mask matrices with the same number of rows as the wire rope bar diagram and 6 columns each. The first grayscale mask matrix is ​​used on the wire rope strip image of the current frame, and the grayscale value of each row is (0, 0.2, 0.4, 0.6, 0.8, 1). The second grayscale mask matrix is ​​used on the wire rope strip image of the previous frame, and the grayscale value of each row is (1, 0.8, 0.6, 0.4, 0.2).

[0063] S5.2. Use the entire first frame of the wire rope strip image as the splicing area to form the starting panoramic image to be spliced.

[0064] S5.3 Starting from the second frame of the wire rope strip image, take the last k columns of the non-stitched area of ​​the current frame of the wire rope strip image and the first k columns of the current panoramic image as the stitching transition area, multiply them by the first and second grayscale mask matrices respectively, and then add them together. The result of the sum replaces the first k columns of the current panoramic image. Finally, add the stitching area of ​​the current frame of the wire rope strip image to the beginning of the current panoramic image.

[0065] S5.4 Repeat step S5.3 until all wire rope strip diagrams are used, thus achieving a panoramic splicing of the wire rope's cylindrical surface. Figure 8 As shown, a panoramic view of the cylindrical surface of the wire rope is obtained.

[0066] Although the illustrative specific embodiments of the present invention have been described above to enable those skilled in the art to understand the invention, it should be understood that the invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the invention as defined and determined by the appended claims, and all inventions utilizing the concept of the present invention are protected.

Claims

1. A method for panoramic splicing of steel wire rope on a cylindrical surface, characterized in that, Includes the following steps: (1) Filming a video of the steel wire rope; (1.1) Set up two white LED light sources. The two white LED light sources are fixed in front of and behind the camera respectively along the axis of the steel wire rope. By adjusting the intensity of the two white LED light sources and the focal length of the camera, the surface texture of the steel wire rope can be displayed in the camera. Then record the distance between the camera and the steel wire rope in this state. (1.2) Keep the distance between the camera and the wire rope constant, rotate the camera around the wire rope at a constant speed for one revolution, and shoot the surface of the wire rope from directly above to obtain the video of the wire rope; (2) Select the effective frames of the wire rope video; (2.1) Binarize each frame of the wire rope video, then use the Canny algorithm to perform edge detection on the binarized image to extract edge pixels. Finally, count the number of edge pixels in each frame, denoted as . , , This indicates the total number of frames in the wire rope video. (2.2) Divide each frame of the wire rope video into several groups of image frames in chronological order. Each group includes Frame image, if the total number of frames no If the number is a multiple of the number of frames remaining, then the last group of numbers will be based on the remaining number of frames. (2.3) For each group of image frames, take the number of pixels with the most edge points. One frame of the image is taken as the valid frame in this set of image frames. ; (3) Obtain the wire rope strip diagram; (3.1) For each set of image frames, extract the corresponding edge-detected image of each valid frame using the Canny algorithm. Treat each frame as a matrix, and then count the number of edge pixels in each column of each matrix to obtain a row of statistical vectors. Let the first row be the edge pixel. The statistical vector corresponding to the valid frames in the group is , ; (3.2) Statistical vector Each element in the set threshold Compare them and find the first one that is greater than or equal to. The elements up to the last one are greater than or equal to The sequence number of the element in the corresponding matrix is ​​used to find the corresponding column position in the corresponding valid frame and perform image segmentation. Then, the image part with the corresponding sequence number is extracted as the pure steel wire rope image of the corresponding valid frame. (3.3) Calculate the average width of all pure steel wire images. Then, crop the center of each pure steel wire rope image. / 3 width steel wire rope strip diagram; (4) Obtain the stitching area on the panoramic image from the steel wire rope strip image; (4.1) Calculate the maximum number of columns to be spliced. ; ； (4.2) Starting from the second wire rope strip diagram, calculate the similarity index between the current wire rope strip diagram and the previous wire rope strip diagram; Then, find the overlapping area with the highest similarity index between the current wire rope strip image and the previous wire rope strip image, and use it as the common part of the two wire rope strip images. The remaining part of the current wire rope strip image is used as the stitching area contributed to the panoramic image. (5) Steel wire rope strip splicing; (5.1) Construct two rows with the same number of rows as the wire rope bar diagram and the same number of columns. The grayscale mask matrix of the column, where, The first grayscale mask matrix is ​​used on the current wire rope strip image, and the grayscale value of each row gradually increases from 0 to 1. The number of rows is determined by the second grayscale mask matrix, which is used on the previous frame's wire rope strip image. The grayscale values ​​in each row gradually decrease from 1 to 0. Number; (5.2) Use the entire first wire rope strip diagram as the splicing area to form the starting panoramic image to be spliced; (5.3) Starting from the second wire rope strip diagram, the last non-spliced ​​area of ​​the current wire rope strip diagram is... The first image in the series and the current panorama The column images are used as a stitching transition area. They are multiplied by the first and second grayscale mask matrices respectively and then added together. The result of the sum replaces the foreground of the current panoramic image. Finally, add the stitching area of ​​the current wire rope strip image to the beginning of the current panoramic image; (5.4) Repeat step (5.3) until all the steel wire rope strip diagrams are used up, thereby realizing the panoramic splicing of the steel wire rope cylindrical surface.

Images