Intelligent Detection Method and System for Belt Conveyor

Abstract

This application relates to the technical field of image data processing, and particularly to an intelligent detection method and system for a belt conveyor. The method includes the steps of: segmenting the conveyor belt image to form multiple tiles; calculating the crack score of each tile; judging the operating state of the conveyor according to the crack score of the tile. The calculation step of the crack score includes: calculating the crack suspicion degree of the tile, and taking the product of the crack suspicion degree and the sample entropy of the tile as the crack degree; determining the crack score based on the crack degree of the tile. This application has the effect of improving the accuracy of conveyor operation detection.

Description

Technical Field

[0001] This application relates to the technical field of image data processing, and particularly to an intelligent detection method and system for a belt conveyor. Background Art

[0002] A belt conveyor is a common transportation device, which is applied in industries such as mines, metallurgy, and coal mining. In order to reduce the occurrence of faults during the operation of the conveyor, it is necessary to detect the conveyor. The detection of the conveyor involves the detection of multiple items, including the detection of longitudinal tearing of the conveyor belt. This detection is directly related to whether the conveyor can transport materials normally and is particularly important in the detection items of the conveyor. The traditional manual detection method has the problems of response lag and low efficiency. Therefore, with the development and progress of information technology and image processing technology, the function of detecting cracks in an object to be detected by collecting images of the object to be detected has been gradually realized. The sample entropy method can be applied to the crack detection of the conveyor belt; when dealing with image problems, it first divides the image into several tiles, then calculates the sample entropy value according to the gray value performance of the pixel points in each tile, and then judges whether there are longitudinal cracks in the object to be detected according to the sample entropy. The application of this method has reduced the risk of accidents during material transportation to a certain extent.

[0003] However, in actual production activities, when materials move on the surface of the conveyor belt, they will rub against the surface of the conveyor belt and form longitudinal drag marks. The drag marks will have a similar gray value performance to potential longitudinal cracks, and the tiles they belong to will have a sample entropy value highly close to that of the tiles where potential longitudinal cracks are located. Therefore, the existence of drag marks causes the above sample entropy-based method to be unable to accurately detect potential longitudinal crack defects; furthermore, it leads to an incorrect judgment of the operating state of the conveyor belt, and finally results in a low detection accuracy of the conveyor. Summary of the Invention

[0004] In order to improve the accuracy of conveyor belt detection, this application provides an intelligent detection method and system for a belt conveyor.

[0005] In a first aspect, this application provides an intelligent detection method for a belt conveyor, adopting the following technical solution:

[0006] An intelligent detection method for a belt conveyor includes the steps of: segmenting a conveyor belt image to form multiple tiles; calculating the crack score of each tile; and judging the operating state of the conveyor according to the crack score of the tile.

[0007] The calculation steps of the crack score include: calculating the crack suspicion degree of the tile, taking the product of the crack suspicion degree and the sample entropy of the tile as the crack degree; and determining the crack score based on the crack degree of the tile. The formula for calculating the crack suspicion degree is: ; where represents the crack suspicion degree of the th tile; represents the gray complexity within the th tile, represents the number of pixel points whose gradient change direction is perpendicular to the longitudinal direction of the conveyor belt within the th tile, represents the variance value of the gradient change directions of all pixel points with gradient changes within the th tile.

[0008] In this application, the crack suspicion degree of the tile is introduced on the basis of sample entropy, enhancing the sensitivity to cracks in the conveyor belt image. The crack suspicion degree is jointly controlled by the gray complexity of the tile, the gradient direction of the pixel points in the tile, and the gradient change of the pixel points in the tile, making the detection of cracks more accurate and sensitive. For the area with cracks, the distribution of gray levels in the tile is more chaotic and complex. Therefore, the gray complexity of the tile can be used to judge the tile with cracks. At the same time, in the tile with longitudinal cracks, there must be pixel points whose gradient direction is perpendicular to the longitudinal direction of the conveyor belt. Therefore, this factor is also an important condition for judging whether there are cracks in the tile. In addition, during the production process of the conveyor belt, in order to improve its own load-bearing capacity, fibrous materials are generally embedded inside the conveyor belt. After cracks occur, the fibrous materials are exposed outside the conveyor belt, causing changes in the gradient directions of each pixel point in the conveyor belt image. Therefore, introducing the variance value of the gradient change directions of the pixel points with gradient changes in the formula can further improve the accuracy of judging whether there are cracks in the tile. Compared with the related technology that only judges cracks in the conveyor belt image through sample entropy, the method in this application considers more comprehensively and has higher accuracy, thus improving the accuracy of conveyor detection.

[0009] Optionally, the calculation steps of the gray complexity of the tile include: calculating the gray diversity of the tile and the maximum gray difference of the tile; taking the product of the gray diversity and the maximum gray difference as the gray complexity of the tile.

[0010] Calculate the gray complexity of the tile from multiple aspects in a coordinated manner, thereby improving the sensitivity to the gray level in the tile and the accuracy of subsequent crack detection.

[0011] Optionally, the calculation steps of the gray diversity of the tile include: calculating the number of different gray values in the dataset composed of the gray values of all pixel points in the tile; calculating the total number of pixel points in the tile, and taking the ratio of the number of different gray values to the total number of pixel points as the gray diversity of the tile.

[0012] Calculate the number of different grayscale values ​​in the tile. The larger the value, the more chaotic the grayscale of each pixel in the tile, which makes the grayscale complexity of the tile greater. Divide this number by the total number of pixels in the tile to standardize the calculation result, which is convenient for comparing the grayscale complexity of multiple tiles.

[0013] Optionally, the steps for calculating the maximum grayscale difference of a tile include: obtaining the grayscale value of the pixel with the largest grayscale value in the tile, and defining the grayscale value as the maximum grayscale value of the tile; obtaining the grayscale value of the pixel with the smallest grayscale value in the tile, and defining the grayscale value as the minimum grayscale value of the tile; setting hyperparameters, and taking the sum of the hyperparameters and the minimum grayscale value as the grayscale sum; and taking the ratio of the maximum grayscale value to the grayscale sum as the maximum grayscale difference of the tile.

[0014] The comparison between the maximum grayscale value of the pixel in the block and the minimum grayscale value reflects the degree of grayscale change in the block. The more drastic the grayscale change in the block, the greater the maximum grayscale difference of the block, and thus the greater the credibility of the grayscale value chaos in the block, thereby improving the accuracy of the grayscale complexity of the block.

[0015] Optionally, the method for segmenting the conveyor belt image comprises the steps of: selecting a segmentation reference pixel point; dividing the image into Each area is a tile.

[0016] Optionally, the reference pixel point is a pixel point at a corner position of the conveyor belt image.

[0017] In the process of conveyor belt image segmentation, there may be pixels in the conveyor belt image that cannot be completely divided into multiple The size of the tile exists at the edge of the conveyor belt image and cannot be divided into For example, the conveyor belt image includes 60 rows and 60 columns of pixels, and the preset block size is . In the conveyor belt image segmentation process, the conveyor belt image can be divided into three blocks in the horizontal and vertical directions, and three columns of pixels and three rows of pixels remain. If the non-corner position of the conveyor belt image is selected as the reference pixel point for segmentation, there may be undivided pixels on the four edges corresponding to the conveyor belt image. Therefore, the reference pixel point is selected at the corner of the conveyor belt image, so that the two edges adjacent to the reference pixel point during the segmentation process will not have pixels that cannot be divided into blocks, which is convenient for processing the conveyor belt image.

[0018] Optionally, the crack score is a normalized value of the crack severity.

[0019] The degree of cracks is normalized to control the value range of the calculation results, which is convenient for subsequent judgment on whether there are cracks in the block.

[0020] Optionally, the method for judging the operation state of the conveyor according to the crack score of the tile includes the steps of: setting a score threshold; judging that the tile has a crack defect when the score of the tile is greater than the score threshold; and judging that the conveyor has an operation failure and issuing an alarm when there is a tile with a crack defect.

[0021] In a second aspect, the present application provides an intelligent detection system for a belt conveyor, adopting the following technical solution:

[0022] An intelligent detection system for a belt conveyor, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent detection method for a belt conveyor as described above is implemented.

[0023] Generate a computer program for the intelligent detection method of the belt conveyor as described above and store it in the memory to be loaded and executed by the processor. Thus, a system is made according to the memory and the processor, which is convenient to use.

[0024] The present application has the following technical effects: By information such as the gray complexity of the tile, the gradient of the pixel points in the tile, and the change in the gradient direction, the crack suspicion degree of the tile is calculated. The crack suspicion degree is combined with the sample entropy of the tile to calculate the crack degree of each tile, so that the detection algorithm can more accurately distinguish longitudinal cracks from drag marks and accurately identify the cracks. Finally, the tiles are judged by the crack score, which improves the accuracy of the final conveyor belt detection result and further improves the accuracy of the conveyor operation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flowchart of an intelligent detection method for a belt conveyor according to an embodiment of the present application.

[0026] Figure 2 is a flowchart of step S2 in an embodiment of the present application. DETAILED DESCRIPTION

[0027] An embodiment of the present application discloses an intelligent detection method for a belt conveyor. After the image of the conveyor belt is segmented, a plurality of tiles are formed, a crack score that can reflect the possibility of cracks in the tile is calculated, and whether there are cracks in the tile is judged according to the crack score.

[0028] Referring to Figure 1 , the intelligent detection method for the belt conveyor includes step S1-step S3.

[0029] Step S1: Segment the conveyor belt image to form a plurality of tiles;

[0030] Use an image acquisition device (such as a camera) to collect images of the surface of the conveyor belt of the conveyor, and perform grayscale processing on the images to obtain conveyor belt images.

[0031] When collecting images of the surface of the conveyor belt through the acquisition device, the direction of the camera should be perpendicular to the belt body to ensure that the belt body is horizontal in the surface image, which is convenient for subsequent processing of the surface image.

[0032] Determine the reference pixel point in the conveyor belt image. Taking this pixel point as the reference (which can also be understood as the starting pixel point), divide the conveyor belt image into multiple regions of size , and each region is a tile.

[0033] In this embodiment, the reference pixel point is the pixel point at the upper left corner of the conveyor belt image. In other embodiments, it can also be the pixel point at the upper right corner, lower left corner, or lower right corner of the conveyor belt image. The reason for choosing the pixel point at the corner of the conveyor belt image as the reference pixel point here is to reduce the situation where the edge of the conveyor belt image is not sufficient to form a region during the segmentation of the conveyor belt image. Referring to the reference pixel point at the upper left corner of the conveyor belt image in this embodiment, then during the subsequent segmentation of the conveyor belt image, the situation where the two edges corresponding to the reference pixel point are not sufficient to be divided into regions of size is reduced, which is convenient for subsequent calculations. In this embodiment, the size of the tile is

[0034] Set a width threshold. During the segmentation of the conveyor belt image, if the edge of the conveyor belt image is not sufficient to be divided into regions of size , obtain the width of the region formed by the remaining pixel points. If this width is less than the width threshold, divide the remaining pixel points into the adjacent regions that have been completed. If this width is greater than the width threshold, divide the remaining pixel points into multiple regions of size equal to this width to form new tiles.

[0035] In this embodiment, the size of the tile is , and the width threshold is 9.

[0036] Step S2: Calculate the crack score of each tile;

[0037] Referring to Figure 2 , step S2 includes steps S21 - S23

[0038] S21: Calculate the crack suspicion degree of the tile, and take the product of the crack suspicion degree and the sample entropy of the tile as the crack degree;

[0039] Calculate the gray complexity of the tile;

[0040] Calculate the gray diversity of the tile and the maximum gray difference of the tile;

[0041] The calculation steps of the gray-scale diversity of a tile include: calculating the number of different gray-scale values in the dataset composed of the gray-scale values of all pixel points in the tile; calculating the total number of pixel points in the tile, and taking the ratio of the number of different gray-scale values to the total number of pixel points as the gray-scale diversity of the tile.

[0042] The calculation steps of the maximum gray-scale difference of a tile include: obtaining the gray-scale value of the pixel point with the largest gray-scale value in the tile and defining this gray-scale value as the maximum gray-scale value of the tile; obtaining the gray-scale value of the pixel point with the smallest gray-scale value in the tile and defining this gray-scale value as the minimum gray-scale value of the tile; setting a hyperparameter, and taking the sum of the hyperparameter and the minimum gray-scale value as the gray-scale sum; taking the ratio of the maximum gray-scale value to the gray-scale sum as the maximum gray-scale difference of the tile.

[0043] In one embodiment, the product of the gray-scale diversity and the maximum gray-scale difference is the gray-scale complexity of the tile.

[0044] Specifically, the calculation formula of the gray-scale complexity of a tile can be expressed as:

[0045] ; where represents the gray-scale complexity within the th tile, the th tile is the number of different gray-scale values in the dataset composed of the gray-scale values of all pixel points in the tile, represents the th number of pixel points within the tile. and represent the maximum and minimum gray-scale values of all pixel points within the th tile. is the hyperparameter. , and its existence is to prevent from taking the value of 0.

[0046] In the formula represents the proportion of the number of different gray-scale values in a tile to the total number of pixel points in the tile. For example, if a tile has four pixel points, and the gray-scale values of the four pixel points are 22, 23, 22, and 22 respectively; it can be seen that the gray-scale values of all pixel points in the tile are divided into two categories, namely 22 and 23; then , the larger this value is, the more chaotic the gray-scale values within a block are, the greater the corresponding degree of gray-scale chaos is, and the greater the possibility that the block belongs to a potential longitudinal crack defect.

[0047] represents the relative size of the maximum gray-scale value to the minimum gray-scale value within each block. In the conveyor belt image, longitudinal crack defects will have a more intense change in gray-scale values. Therefore The larger it is, the greater the possibility of cracks in the tile, and the greater the corresponding gray complexity.

[0048] In one embodiment, the gray complexity of the tile can also be the sum of the gray diversity and the maximum gray difference.

[0049] Calculate the crack suspicion degree;

[0050] The calculation formula for the crack suspicion degree is:

[0051] ; where, where, represents the crack suspicion degree of the th tile; represents the gray complexity within the th tile, represents the number of pixel points whose gradient change direction is perpendicular to the longitudinal direction of the conveyor belt within the th tile, represents the variance value of the gradient change directions of all pixel points with gradient changes within the th tile.

[0052] In the formula, for tiles with no gradient change, the corresponding and do not exist. In this case, let and both be 1.

[0053] The crack suspicion degree is used to represent the possibility of cracks in the tile. In the formula, represents the number of pixel points whose gradient direction is perpendicular to the longitudinal direction of the conveyor belt in the tile. If is larger in a tile, it indicates that the texture in the tile is more likely to have the characteristic of longitudinal extension, and thus the corresponding crack suspicion degree is larger.

[0054] During the production and manufacturing process of the conveyor belt, in order to improve the load-bearing capacity of the conveyor belt itself, fiber materials are usually embedded inside it. When cracks appear in the conveyor belt, such fibers will be exposed outside the conveyor belt, forming unique characteristics on the surface of the conveyor belt. Therefore, this feature can be used to further judge the crack suspicion degree of the tile. In the formula, represents the variance value of the gradient direction change value. The larger this value is, the more pixel points with inconsistent gradient change directions in the tile, which further indicates that the possibility of different-direction textures in the tile is greater (corresponding to the fibrous substances exposed outside the conveyor belt), and finally the crack suspicion degree of the tile is greater.

[0055] S22: Calculate the crack degree; the crack degree is the product of the crack suspicion degree and the sample entropy of the tile.

[0056] The calculation formula for the crack degree can be expressed as:

[0057] ; where, in the formula represents the crack degree of the th sub-block; represents the crack suspicion degree of the th tile; represents the sample entropy of the th tile. The calculation of sample entropy is common knowledge to those skilled in the art and will not be elaborated here.

[0058] S23: Determine the crack score based on the crack degree of the tile;

[0059] The crack score of the tile is the value after normalizing the crack degree of the tile. In this embodiment, the function is used to normalize the crack degree.

[0060] Step S3: Judge the running state of the conveyor according to the crack score of the tile;

[0061] Set a score threshold; when the score of the tile is greater than the score threshold, it is judged that the tile has a crack defect; when there is a tile with a crack defect, it is determined that the conveyor has a running fault and an alarm is issued.

[0062] The embodiment of the present application also discloses an intelligent detection system for a belt conveyor, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent detection method for a belt conveyor according to the present application is implemented.

[0063] The above system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface, and their settings and functions are known in the art, so they will not be elaborated here.

[0064] The above are all preferred embodiments of the present application. The protection scope of the present application is not limited by this. Therefore, all equivalent changes made according to the structure, shape, and principle of the present application should be covered within the protection scope of the present application.

Claims

1. An intelligent detection method for a belt conveyor, characterized in that, It includes steps: segmenting the conveyor belt image to form multiple tiles; calculating the crack score of each tile; judging the running state of the conveyor according to the crack score of the tile; The calculation steps of crack scoring include: calculating the crack suspicion degree of the tile, taking the product of the crack suspicion degree and the sample entropy of the tile as the crack degree; determining the crack score based on the crack degree of the tile; where the calculation formula for the crack suspicion degree is: ; in the formula, represents the crack suspicion degree of the th tile; represents the gray complexity within the th tile, represents the number of pixel points within the th tile whose gradient change direction is perpendicular to the longitudinal direction of the conveyor belt, represents the variance value of the gradient change directions of all pixel points with gradient changes within the th tile; The calculation steps of the grayscale complexity of a tile include: calculating the grayscale diversity of the tile and the maximum grayscale difference of the tile; taking the product of the grayscale diversity and the maximum grayscale difference as the grayscale complexity of the tile, and the calculation formula of the grayscale complexity of the tile can be expressed as: ; In the formula, represents the grayscale complexity within the -th tile, the number of different grayscale values in the dataset composed of the grayscale values of all pixel points in the -th tile, represents the number of pixel points within the -th tile, and represent the maximum grayscale value and the minimum grayscale value of all pixel points within the -th tile, .

2. The intelligent detection method of a belt conveyor according to claim 1, wherein The calculation steps of the gray diversity of the tile include: calculating the number of different gray values in the dataset composed of the gray values of all pixel points in the tile; calculating the total number of pixel points in the tile, and taking the ratio of the number of different gray values to the total number of pixel points as the gray diversity of the tile.

3. The intelligent detection method of a belt conveyor according to claim 1, characterized in that, The calculation steps of the maximum gray difference of the tile include: obtaining the gray value of the pixel point with the largest gray value in the tile and defining this gray value as the maximum gray value of the tile; obtaining the gray value of the pixel point with the smallest gray value in the tile and defining this gray value as the minimum gray value of the tile; setting a hyperparameter, and taking the sum of the hyperparameter and the minimum gray value as the gray sum; taking the ratio of the maximum gray value to the gray sum as the maximum gray difference of the tile.

4. The intelligent detection method of a belt conveyor according to claim 1, characterized in that, The method for segmenting the conveyor belt image includes the steps of: selecting a segmentation reference pixel point; starting from the reference pixel point, dividing regions with a size of Each region is a tile.

5. The intelligent detection method of a belt conveyor according to claim 4, wherein, The reference pixel point is the pixel point at the corner position of the conveyor belt image.

6. The intelligent detection method of a belt conveyor according to claim 4, characterized in that When there are not enough remaining pixels at the edge of the conveyor belt image to be divided into tiles of the size, calculate the width of the area formed by the remaining pixel points; when the width is less than the preset width threshold, divide the remaining pixel points into the adjacent areas that have been completed; when the width is greater than the preset threshold, divide the remaining pixel points into a new area.

7. An intelligent detection method for a belt conveyor according to claim 1, characterized in that, The crack score is the value after normalizing the crack degree.

8. An intelligent detection method for a belt conveyor according to claim 1, characterized in that The method for judging the running state of the conveyor according to the crack score of the tile includes steps: setting a score threshold; judging that the tile has a crack defect when the score of the tile is greater than the score threshold; determining that the conveyor has a running fault and issuing an alarm when there is a tile with a crack defect.

9. An intelligent detection system for a belt conveyor, characterized in that, It includes: a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent detection method for a belt conveyor according to any one of claims 1-8 is implemented.

Images