Conveyor belt edge damage assessment method based on image mask difference calculation

Abstract

The invention discloses a conveyor belt edge damage assessment method based on image mask difference calculation, which relates to the technical field of automatic detection, and comprises the following steps of: acquiring an image of a conveyor belt in real time, processing the real-time image by adopting an image instance segmentation algorithm, and respectively extracting a normal reference mask and an edge damage mask; according to the method, an edge damage area is quantified by calculating a difference value between a normal reference mask and an edge damage mask in the edge area, the severity of the damage of the edge area of the conveyor belt is evaluated according to accumulated damage, damage of different levels is generated for the conveyor belt, and meanwhile corresponding treatment measures are preset for the damage of the different levels. Compared with traditional manual inspection and a high-cost sensor technology, detection is achieved through a low-cost camera and a corresponding algorithm, the implementation cost is lower, and meanwhile the real-time performance and accuracy of detection are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of automated detection, and specifically, to a method for evaluating conveyor belt edge damage based on image mask difference calculation. Background Art

[0002] Conveyor belts are widely used in industrial production, especially in fields such as logistics, mining, and metallurgy. However, edge damage to the conveyor belt can lead to a decline in its performance, and in severe cases, it can even affect the normal operation of the production line. Currently, the detection of conveyor belt edge damage mainly relies on manual inspections or detections based on simple sensors. Although it can meet certain detection functions, there are still the following problems: 1) Poor real-time performance. Sensor data usually cannot be processed and fed back in real time, making it difficult to immediately evaluate the damage to the conveyor belt; 2) Low accuracy. Manual detection is greatly affected by subjective factors, and sensors cannot distinguish different types of damage, resulting in a high rate of false positives and false negatives, that is, it is difficult for the existing technology to accurately identify the location and degree of conveyor belt edge damage, and it is easy to miss detections or misdetect; 3) The existing technology relies on manual labor or expensive sensor equipment, resulting in high detection costs. Especially in large production lines, the costs of manual inspections and equipment maintenance cannot be ignored.

[0003] Therefore, there is an urgent need for a method for detecting conveyor belt edge damage to solve the problems of low real-time performance and accuracy, and high detection costs in the detection of conveyor belt edge damage in the existing technology. Summary of the Invention

[0004] To solve the deficiencies of the above-mentioned existing technology, the present invention provides a method for evaluating conveyor belt edge damage based on image mask difference calculation, including:

[0005] Collect real-time images of the conveyor belt, where the acquisition area of the real-time images includes the edge areas on both sides of the conveyor belt, and the edge area is an area with a preset width from the side of the conveyor belt;

[0006] Process all real-time images using a preset algorithm, divide the edge area into a normal area and a damaged area, and extract the images of all normal areas and the images of all damaged areas respectively. Based on the images of all normal areas, generate a normal reference mask, and based on the images of all damaged areas, generate an edge damage mask;

[0007] Perform equidistant transverse cutting on the conveyor belt to obtain a number of transverse ends including the edge areas on both sides of the conveyor belt;

[0008] Calculate the pixel difference value of the edge area in each transverse segment using a first formula to obtain a first calculation result;

[0009] Based on the first calculation result, use a second formula to accumulate and calculate the pixel difference values in the edge regions of all the horizontal segments on the conveyor belt, obtaining a second calculation result;

[0010] Based on the second calculation result, use a third formula to calculate the damage degree value of the edge region on the conveyor belt;

[0011] Based on the damage degree value of the edge region on the conveyor belt, classify the damage degree of the edge region on the conveyor belt, and preset corresponding treatment measures for different levels.

[0012] The present invention is implemented through the following technical solutions: In this solution, the image of the conveyor belt is collected in real time, and the collection area needs to ensure that it covers the edge region of the conveyor belt. The edge region of the conveyor belt is the region at a preset width (10 - 20 cm) from the side of the conveyor belt; Use a preset algorithm (image instance segmentation algorithm) to process the collected conveyor belt image, extract the normal region and the damaged region in the edge region of the conveyor belt respectively, and then generate a normal reference mask for the normal region and an edge damage mask for the damaged region respectively; By calculating the difference value between the normal reference mask and the edge damage mask in the edge region, quantify the difference between the normal region and the damaged region, and then accumulate the difference values between the normal region and the damaged region to obtain an accumulation result. Generate the damage degree value of the edge region according to the accumulation result, and use the damage degree value to evaluate the severity of the damage to the edge region of the conveyor belt. Specifically, classify the damage to the edge region of the conveyor belt based on the loss degree value, and preset corresponding treatment measures for different levels of damage.

[0013] As an alternative technical solution, classifying the damage degree of the edge region on the conveyor belt and presetting corresponding treatment measures for different levels includes:

[0014] When the damage degree value of the edge region of the conveyor belt is between 0% and 20%, determine that the conveyor belt has a first-level damage, and conduct regular inspections on the conveyor belt;

[0015] When the damage degree value of the edge region of the conveyor belt is between 21% and 50%, determine that the conveyor belt has a second-level damage, and conduct regular monitoring and evaluation on the conveyor belt to obtain an evaluation result; Based on the evaluation result, maintain the conveyor belt;

[0016] When the damage degree value of the edge region of the conveyor belt is between 51% and 100%, determine that the conveyor belt has a third-level damage, and conduct shutdown maintenance on the conveyor belt.

[0017] As an alternative technical solution, the first formula includes:

[0018]

[0019] ΔD row is the pixel difference value of the edge region in each horizontal segment, and m is the number of pixels included in the edge region in each horizontal segment; P normal，j is the value of the j-th pixel in the normal reference mask, indicating whether the corresponding position belongs to the normal region; P damage，j is the value of the j-th pixel in the edge damage mask, indicating whether the corresponding position belongs to the damaged region.

[0020] As an optional technical solution, the second formula includes:

[0021]

[0022] ΔD total is the pixel difference value of the edge region on the conveyor belt, and k is the number of the horizontal segments.

[0023] As an optional technical solution, the third formula includes:

[0024]

[0025] S is the damage degree value of the edge region of the conveyor belt, D max is the theoretical maximum value of the pixel difference value of the edge region on the conveyor belt.

[0026] As an optional technical solution, obtaining the theoretical maximum value D of the pixel difference value of the edge region on the conveyor belt max includes:

[0027] Based on the size of the conveyor belt, the resolution of the real-time image, the pixel distribution in the normal reference mask, and the pixel distribution in the edge damage mask, the theoretical maximum D is preset max .

[0028] As an optional technical solution, processing all real-time images by using a preset algorithm includes: the preset algorithm is an image instance segmentation algorithm.

[0029] As an optional technical solution, an industrial camera is used to collect the real-time image of the conveyor belt.

[0030] One or more technical solutions provided by the present invention have at least the following technical effects or advantages:

[0031] The present invention discloses a method for evaluating the edge damage of a conveyor belt based on image mask difference calculation. By collecting the images of the conveyor belt in real time and using an image instance segmentation algorithm to process the real-time images, a normal reference mask and an edge damage mask are respectively extracted. By calculating the difference value between the normal reference mask and the edge damage mask in the edge area, the edge damage area is quantified, and the severity of the damage in the edge area of the conveyor belt is evaluated according to the cumulative damage, generating different levels of damage to the conveyor belt. At the same time, corresponding treatment measures are preset for different levels of damage. Through real-time image collection and image processing, the present invention can realize real-time monitoring of the edge area of the conveyor belt. Through the instance segmentation algorithm and mask difference calculation, it can accurately identify and quantify the degree of edge damage. Compared with traditional manual inspection and high-cost sensor technologies, the present invention realizes detection through a low-cost camera and corresponding algorithms, with lower implementation costs, while ensuring the real-time nature and accuracy of the detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, and constitute a part of the present invention, but do not limit the embodiments of the present invention.

[0033] Figure 1 It is a schematic flowchart of a method for evaluating the edge damage of a conveyor belt based on image mask difference calculation in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other.

[0035] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0036] Embodiment

[0037] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a method for evaluating the edge damage of a conveyor belt based on image mask difference calculation in the present invention, including:

[0038] Collect real-time images of the conveyor belt. The acquisition area of the real-time images includes the edge areas on both sides of the conveyor belt, and the edge area is an area with a preset width from the side of the conveyor belt.

[0039] Using a preset algorithm to process all real-time images, the edge area is divided into a normal area and a damaged area, and images of all normal areas and images of all damaged areas are extracted respectively, a normal reference mask is generated based on the images of all normal areas, and an edge damage mask is generated based on the images of all damaged areas;

[0040] Cutting the conveyor belt into equal-distance transverse sections to obtain a plurality of transverse ends including edge regions on both sides of the conveyor belt;

[0041] Calculating the pixel difference value of the edge area in each horizontal segment using the first formula to obtain a first calculation result;

[0042] Based on the first calculation result, a second formula is used to accumulate and calculate the pixel difference values of the edge areas in all the transverse sections on the conveyor belt to obtain a second calculation result;

[0043] Based on the second calculation result, a third formula is used to calculate the damage degree value of the upper edge area of the conveyor belt;

[0044] Based on the damage degree value of the upper edge area of the conveyor belt, the damage degree of the upper edge area of the conveyor belt is divided into levels, and corresponding treatment measures are preset for different levels.

[0045] The specific embodiments of the present invention are as follows:

[0046] Image acquisition:

[0047] Real-time images of the conveyor belt are captured by a high-resolution industrial camera. In order to ensure that the image resolution and acquisition frequency meet the detection requirements, it is recommended to use a camera with a resolution of not less than 1920X1080 pixels or higher, and the frame rate of the camera should be not less than 30 frames per second, so as to capture the damage changes in the edge area of the conveyor belt in real time. In addition to using high-resolution industrial cameras, image acquisition equipment can also use machine vision equipment. The specific model can be selected according to the size of the conveyor belt, working environment and actual needs; and regarding the acquisition area, when acquiring real-time images, it should be ensured that the edge area of the conveyor belt is covered (usually an area of 10 to 20 cm wide, and the specific width can be adjusted according to the actual size of the conveyor belt). In addition, the image acquisition angle should try to avoid excessive tilt to ensure that the edge of the conveyor belt in the image is clearly visible.

[0048] Image instance segmentation: Use an image instance segmentation algorithm to process the real-time image of the conveyor belt, extract the normal area and damaged area in the edge area of the conveyor belt, and generate a normal reference mask and a damaged edge mask respectively. Different instance segmentation models can be selected according to different application scenarios for the image instance segmentation algorithm, such as Mask R-CNN, MaskFormer, DeepLab, etc.; Regarding the normal reference mask, it is obtained by taking an image of the conveyor belt without damage. The collected image should cover the entire edge area of the conveyor belt and ensure that there are no sundries or interfering objects in the image; Regarding the edge damage mask, it is obtained by taking an image of the damaged edge of the conveyor belt; Since the damage in the edge area of the conveyor belt may be distributed along the entire length of the conveyor belt, the taken image should cover the entire length of the conveyor belt or select a representative damaged area for image instance segmentation.

[0049] Mask difference calculation: Horizontally divide the conveyor belt at equal intervals to obtain several horizontal ends containing the edge areas on both sides of the conveyor belt;

[0050] Use the first formula to calculate the pixel difference value in the edge area of each horizontal segment to obtain the first calculation result; The first formula includes:

[0051]

[0052] ΔD row is the pixel difference value in the edge area of each horizontal segment, m is the number of pixels included in the edge area of each horizontal segment; P normal，j is the value of the j-th pixel point in the normal reference mask, indicating whether the corresponding position belongs to the normal area; P damage，j is the value of the j-th pixel point in the edge damage mask, indicating whether the corresponding position belongs to the damaged area;

[0053] Based on the first calculation result, use the second formula to accumulate and calculate the pixel difference values in the edge areas of all horizontal segments on the conveyor belt to obtain the second calculation result; The second formula includes:

[0054]

[0055] ΔD total is the pixel difference value in the edge area of the conveyor belt, k is the number of horizontal segments;

[0056] Based on the second calculation result, use the third formula to calculate the damage degree value of the edge area on the conveyor belt; The third formula includes:

[0057]

[0058] S is the damage degree value of the edge area of the conveyor belt, Dmax is the theoretical maximum value of the pixel difference value in the upper edge area of the conveyor belt.

[0059] Among them, obtaining the theoretical maximum value D of the pixel difference value in the upper edge area of the conveyor belt max includes:

[0060] Based on the size of the conveyor belt, the resolution of the real-time image, the pixel distribution in the normal reference mask, and the pixel distribution in the edge damage mask, preset the theoretical maximum D max .

[0061] By calculating the difference value between the normal reference mask and the edge damage mask, quantify the difference between the normal area and the damaged area, so as to evaluate the degree of edge damage of the conveyor belt.

[0062] Damage assessment: Based on the damage degree value of the upper edge area of the conveyor belt, classify the damage degree of the upper edge area of the conveyor belt, and preset corresponding treatment measures for different levels, specifically including:

[0063] When the damage degree value of the conveyor belt edge area is between 0% and 20%, determine that the conveyor belt is a first-level damage, and conduct regular inspections on the conveyor belt;

[0064] When the damage degree value of the conveyor belt edge area is between 21% and 50%, determine that the conveyor belt is a second-level damage, and conduct regular monitoring and evaluation on the conveyor belt to obtain an evaluation result; Based on the evaluation result, maintain the conveyor belt;

[0065] When the damage degree value of the conveyor belt edge area is between 51% and 100%, determine that the conveyor belt is a third-level damage, and conduct shutdown maintenance on the conveyor belt.

[0066] In this embodiment, through real-time image acquisition and image processing, real-time monitoring of the edge area of the conveyor belt can be realized. Through instance segmentation algorithm and mask difference calculation, the degree of edge damage can be accurately identified and quantified. Compared with traditional manual inspection and high-cost sensor technology, the present invention realizes detection through low-cost cameras and corresponding algorithms, with lower implementation costs, and at the same time ensures the real-time performance and accuracy of detection.

[0067] Furthermore, due to the working characteristics of the conveyor belt (continuous operation and uneven stress), it is prone to cause bilateral symmetrical damage to the conveyor belt (i.e., on the same straight line, there are edge damages on both sides) and unilateral damage (on the same straight line, there is only edge damage on one side). If the edge area of the conveyor belt has unilateral damage, it is easy to cause uneven stress on the conveyor belt, resulting in deviation, and accelerating the wear of the idlers and drums; if the edge area of the conveyor belt has bilateral symmetrical damage, it will reduce the overall load-bearing capacity of the conveyor belt, may cause sudden breakage, resulting in a full-line shutdown and even safety accidents. In summary, it can be seen that the bilateral symmetrical damage to the edge area of the conveyor belt is more harmful. Therefore, this embodiment is further optimized on the basis of the existing solution, specifically including:

[0068] Obtain a pair of arbitrarily symmetrical horizontal segments on both sides of the conveyor belt, denoted as the first horizontal segment and the second horizontal segment respectively. The first horizontal segment and the second horizontal segment are located on the same horizontal line;

[0069] Calculate the difference value between the normal part and the damaged part in the edge area of the horizontal segments on both sides through the first formula, and obtain the first difference value corresponding to the first horizontal segment as a, and the second difference value corresponding to the second horizontal segment as b;

[0070] Since there will be some wear in the operation of the conveyor belt, a first set value is preset in advance. Based on the first set value, it is judged whether the first horizontal segment or the second horizontal segment has edge damage, that is, when the first difference value or the second difference value is lower than the first set value, it is judged that there is no edge damage in the corresponding horizontal segment;

[0071] When it is judged that only one of the first horizontal segment and the second horizontal segment has edge damage, it is determined that this group of horizontal segments has unilateral conveyor belt edge damage, and the first formula is optimized to obtain:

[0072]

[0073] Among them, λ1 is the first coefficient when the horizontal segment has unilateral edge damage;

[0074] When it is judged that both the first horizontal segment and the second horizontal segment have edge damage, that is, the edge area of the conveyor belt has bilateral symmetrical damage, the following formula is used to calculate the difference degree between the first difference value and the second difference value:

[0075]

[0076] Among them, C is the difference degree between the corresponding difference values when both the first horizontal segment and the second horizontal segment have edge damage, that is, the difference degree of the edge area damage degree between the first horizontal segment and the second horizontal segment;

[0077] Set a second set value. When C is less than the second set value, it indicates that the difference in the damage degree of the edge regions of the first horizontal section and the second horizontal section is small, and it is determined that there is symmetric damage on both sides of the conveyor belt. When C is greater than or equal to the second set value, it indicates that the difference in the damage degree of the edge regions of the first horizontal section and the second horizontal section is large, and there is a risk of symmetric damage on both sides of the conveyor belt in the future.

[0078] When it is determined that there is symmetric damage on both sides of the conveyor belt, set a third set value, and determine whether both the first difference value and the second difference value are greater than the third set value. If so, immediately send an alarm message to a preset terminal and perform shutdown maintenance. If not, optimize the first formula to obtain:

[0079]

[0080] Among them, λ2 is the second coefficient when it is determined that there is bilateral edge symmetric damage in the horizontal section.

[0081] When it is determined that there is a risk of symmetric damage on both sides of the conveyor belt, optimize the first formula to obtain:

[0082]

[0083] Among them, λ3 is the third coefficient when it is determined that there is a risk of bilateral edge symmetric damage in the horizontal section.

[0084] It should be noted that λ2 > λ3 > λ1, that is, the severity of the symmetric damage on both sides of the conveyor belt that has been determined > the severity of the risk of symmetric damage on both sides of the conveyor belt > the severity of the unilateral damage of the conveyor belt. The subsequent processing method is similar to the original scheme and will not be specifically elaborated. In this embodiment, on the basis of the existing technical scheme, the scheme is optimized to further improve the accuracy and real-time performance of the detection of the edge region of the conveyor belt.

[0085] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the appended claims are intended to be interpreted to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0086] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A method for evaluating the edge damage of a conveyor belt based on image mask difference calculation, characterized in that Including: Collecting real-time images of the conveyor belt, where the acquisition area of the real-time images includes the edge areas on both sides of the conveyor belt, and the edge area is an area with a preset width from the side of the conveyor belt; Processing all real-time images using a preset algorithm, dividing the edge area into a normal area and a damaged area, respectively extracting the images of all normal areas and the images of all damaged areas, generating a normal reference mask based on the images of all normal areas, and generating an edge damage mask based on the images of all damaged areas; Making equidistant horizontal cuts to the conveyor belt to obtain a number of horizontal ends including the edge areas on both sides of the conveyor belt; Calculating the pixel difference value of the edge area in each horizontal segment using a first formula to obtain a first calculation result; Based on the first calculation result, using a second formula to cumulatively calculate the pixel difference values of the edge areas in all horizontal segments on the conveyor belt to obtain a second calculation result; Based on the second calculation result, using a third formula to calculate the damage degree value of the edge area on the conveyor belt; Based on the damage degree value of the edge area on the conveyor belt, classifying the damage degree of the edge area on the conveyor belt and presetting corresponding treatment measures for different levels.

2. The conveyor belt edge damage assessment method based on image mask difference calculation according to claim 1, wherein Classifying the damage degree of the edge area on the conveyor belt and presetting corresponding treatment measures for different levels includes: When the damage degree value of the edge area of the conveyor belt is between 0 and 20%, determining that the conveyor belt is a first-level damage and conducting regular inspections on the conveyor belt; When the damage degree value of the edge area of the conveyor belt is between 21% and 50%, determining that the conveyor belt is a second-level damage, conducting regular monitoring and evaluation on the conveyor belt to obtain an evaluation result; based on the evaluation result, maintaining the conveyor belt; When the damage degree value of the edge area of the conveyor belt is between 51% and 100%, determining that the conveyor belt is a third-level damage and shutting down the conveyor belt for maintenance.

3. The conveyor belt edge damage assessment method based on image mask difference calculation according to claim 1, characterized in that The first formula includes: ΔD row is the pixel difference value of the edge region in each horizontal segment, and m is the number of pixels included in the edge region in each horizontal segment; P normal，j is the value of the j-th pixel in the normal reference mask, indicating whether the corresponding position belongs to the normal region; P damage，j is the value of the j-th pixel in the edge damage mask, indicating whether the corresponding position belongs to the damage region.

4. A method for evaluating conveyor belt edge damage based on image mask difference calculation according to claim 1, characterized in that, The second formula includes: ΔD total is the pixel difference value of the upper edge area of the conveyor belt, and k is the number of the horizontal segments.

5. A method for evaluating the edge damage of a conveyor belt based on image mask difference calculation according to claim 1, characterized in that, The third formula includes: S is the damage degree value of the edge area of the conveyor belt, D max is the theoretical maximum value of the pixel difference value in the upper edge area of the conveyor belt.

6. The conveyor belt edge damage assessment method based on image mask difference calculation according to claim 5, characterized in that Obtain the theoretical maximum value D of the pixel difference value in the upper edge area of the conveyor belt max including: Preset the theoretical maximum D in advance based on the size of the conveyor belt, the resolution of the real-time image, the pixel distribution in the normal reference mask, and the pixel distribution in the edge damage mask max .

7. A method for evaluating the edge damage of a conveyor belt based on image mask difference calculation according to claim 1, characterized in that, Processing all real-time images using a preset algorithm includes: the preset algorithm is an image instance segmentation algorithm.

8. A method for evaluating conveyor belt edge damage based on image mask difference calculation according to claim 1, characterized in that Collecting the real-time images of the conveyor belt using an industrial camera.

Images