Start-stop two-way linkage machine vision detection method for longitudinal tearing of conveying belt

Abstract

The invention discloses a start-stop bidirectional linkage conveyor belt longitudinal tearing machine vision detection method, and relates to the technical field of coal-electricity conveyor belt safety operation and maintenance, and the method comprises the steps: obtaining a bottom longitudinal tearing condition feature image of a conveyor belt in a running state; frame-by-frame image extraction, image filtering, deformation detection, image relevance decision and mask processing are carried out on the bottom longitudinal tearing condition feature image, an alarm signal and a bottom longitudinal tearing condition feature image after abnormal processing are generated, and the alarm signal and the bottom longitudinal tearing condition feature image after abnormal processing are acquired; system linkage is realized; according to the method, after system deployment and debugging are completed, an alarm threshold value is set in advance according to the running speed of the belt, all-weather continuous image recognition is conducted, automatic grading alarm is conducted when abnormal features of the surface of the belt are detected, screenshots are recorded in a database, and long-time training is not needed; the method is good in image effect, high in tearing recognition rate, high in algorithm accuracy and capable of being stably used for a long time.

Description

Technical Field

[0001] The present invention relates to the technical field of safe operation and maintenance of coal power conveyor belts, and in particular to a machine vision detection method for longitudinal tears in conveyor belts with start-stop bidirectional linkage. Background Art

[0002] Coal-fired power generation conveyor belts are essential equipment in the coal-fired power generation industry, playing a key role in transporting coal from mining sites to processing or power plants. Coal-fired power generation conveyor belts are typically made of high-strength materials such as rubber, plastic, nylon, and polyester, offering excellent wear, tensile, and impact resistance. These properties enable them to withstand the weight of bulk materials like coal and maintain stability during long-distance, high-speed transportation. In underground environments, coal-fired power generation conveyor belts must also meet specific flame retardant and anti-static standards to ensure safe production. The operating principle of a coal-fired power generation conveyor belt relies on its drive system, which includes components such as a motor, reducer, and rollers. These components work together to drive the conveyor belt in continuous motion, moving the coal along the belt surface. This smooth and continuous operation enables efficient coal transportation, significantly reducing labor costs and improving production efficiency. Furthermore, coal-fired power generation conveyor belts can be customized to meet specific requirements, such as adjusting belt width and load capacity, to meet different working environments and usage requirements. In the coal-fired power generation industry, the selection and application of coal-fired power generation conveyor belts are crucial for ensuring production safety and improving economic efficiency. In summary, the coal-fired power conveyor belt is an important means of transportation in the coal-fired power industry. With its high efficiency, stability and safety, it provides strong support for coal mining, transportation and utilization.

[0003] As an important energy transportation machinery product in the coal-fired power generation process, the conveyor belt itself does not have electronic and information status data, but it may produce a variety of abnormal conditions during operation. With the large-scale transportation system and the remote operation of the conveyor belt, the collection and utilization of the status information generated by the conveyor belt during operation has become increasingly important.

[0004] The risk of conveyor belt tears is relatively hidden. In most cases, the conveyor belt tears during the conveying process due to uneven force, foreign objects blocking the conveyor belt, etc., which can cause huge economic losses. Therefore, such abnormal failures require rapid response to prevent further loss of life and property. Currently, traditional conveyor belts require manual inspections during operation to closely monitor the operation of the conveyor belt. However, for ultra-long conveyor belts such as those over 3km long, the dim lighting on site and the inconsistent manual judgment standards make manual inspections extremely difficult. The cost of manual maintenance alone is too high and inefficient. A scientific and efficient detection method is urgently needed to improve the operational safety of such conveyor belts.

[0005] To date, most research on conveyor belt protection has focused on longitudinal tear detection, but it is unable to judge the severity of longitudinal tearing in the belt, and thus is unable to effectively respond in the early, middle and final stages of longitudinal tearing. At the same time, modern large-scale conveyor belt systems have complete automated control systems, but the various tear detection methods are rarely linked with the automated control system, and are unable to effectively guide users to judge the degree of tearing, freeing up users' work attention, and thus unable to further carry out targeted and precise maintenance. Summary of the Invention

[0006] The purpose of the present invention is to provide a start-stop bidirectional linkage conveyor belt longitudinal tear machine vision detection method to solve the above-mentioned shortcomings in the prior art.

[0007] In order to achieve the above-mentioned object, the present invention provides the following technical solution: a start-stop bidirectional linkage conveyor belt longitudinal tear machine vision detection method, comprising:

[0008] S1. Acquire a characteristic image of longitudinal tearing at the bottom of the conveyor belt when it is running;

[0009] S2. Performing frame-by-frame image extraction, image filtering, deformation detection, image correlation decision-making, and mask processing on the bottom longitudinal tearing condition characteristic image to generate an alarm signal and an abnormality-processed bottom longitudinal tearing condition characteristic image;

[0010] S3, obtaining the alarm signal and the characteristic image of the bottom longitudinal tearing condition after abnormality processing, and realizing system linkage;

[0011] S2 includes:

[0012] S21, the image extraction includes cropping and shielding complex invalid pixels on the scene, setting the feature amplification area to algorithm RO I, and binarizing the laser morphology within the algorithm RO I to obtain image extraction data;

[0013] S22, the image filtering includes image distortion correction and image denoising to obtain steady-state feature image data;

[0014] S23, performing sudden deformation edge detection on the steady-state feature image data to obtain sudden deformation data;

[0015] S24. Perform image correlation processing on the sudden deformation data to obtain an abnormality judgment representing a risk level, and output an abnormality feature classification analysis conclusion based on the abnormality judgment.

[0016] Furthermore, the S1 includes:

[0017] S11: After the conveyor belt is running at a uniform speed, the bottom of the conveyor belt is continuously scanned without stroboscopic light, and a real-time image of the bottom of the conveyor belt is collected immediately after the conveyor belt is connected to the material trough for unloading;

[0018] S12: performing frame-by-frame processing and continuous decision-making calculation on the bottom real-time image.

[0019] Furthermore, the S1 further includes:

[0020] S13. Adjust according to the on-site situation to obtain a real-time image of the bottom after correcting and amplifying the longitudinal tearing characteristics.

[0021] Furthermore, the binarization process includes:

[0022]

[0023] Among them, maxval is the maximum value, src is the grayscale image, thresh is the threshold, and the result data dst is the binarized image.

[0024] Furthermore, the image distortion correction includes:

[0025] xdistorted=x(1+k1r2+k2r4+k3r6);

[0026] ydistorted=y(1+k1r2+k2r4+k3r6);

[0027] x and y represent pixels in different distortion directions, r represents the distortion radius, and k represents the distortion coefficient at different levels of visual center distance.

[0028] Furthermore, the image denoising includes:

[0029] The image extraction data is smoothed using a convolution filter and then processed by the cv2.GaussianBlur function of OpenCV, including:

[0030] Smoothed image=cv2.GaussinaBlur(image,(x,y),ksize);

[0031] image represents the full frame pixels of the input image, x and y represent the filter size, and ksize represents the smoothing strength.

[0032] Furthermore, the sudden deformation edge detection uses gradient calculation to detect the direction and degree of data change within the pixel and locate the area where the feature target in the image changes sharply. The gradient calculation is implemented using the sobel operator of OpenCV, which specifically includes:

[0033] Sobel x=cv2.Sobel(Smoothed image,cv2.CV 64F,1,0,ksize);

[0034] Sobel y=cv2.Sobel(Smoothed image,cv2.CV 64F,1,0,ksize);

[0035] Smoothed image represents the steady-state feature image data. 1,0 and 0,1 are used to calculate the gradients in the x-direction and y-direction respectively. ksize represents the gradient calculation strength.

[0036] Furthermore, the image correlation processing performs correlation binding on continuous frames and continuous pixel clusters, and then sets gradient detection thresholds in the associated x-direction and y-direction.

[0037] Furthermore, the S2 further includes:

[0038] S25. Generate a corresponding alarm signal based on the abnormal feature hierarchical analysis conclusion, perform mask processing on the bottom longitudinal tearing condition characteristic image, mark the abnormality, obtain the bottom longitudinal tearing condition characteristic image after the abnormality processing, and send the bottom longitudinal tearing condition characteristic image after the abnormality processing to the user terminal.

[0039] Furthermore, the S3 includes:

[0040] S31, realize two-way real-time communication with the main coal power management system platform;

[0041] S32, control the conveyor belt longitudinal tear machine vision detection to be consistent with the conveyor belt start and stop status;

[0042] S33: Send a prompt instruction to the user terminal according to the alarm signal, and receive an operation instruction from the user terminal.

[0043] Compared with the existing technology, the present invention provides a start-stop bidirectional linkage conveyor belt longitudinal tear machine vision detection method:

[0044] 1. Automatic feature determination of digital images based on MATLAB. It mainly highlights the image features of conveyor belt tears through the combination of laser and light source. High-resolution and high-frame-rate industrial cameras are used for frame-by-frame acquisition to obtain feature images that are easy for both machine vision algorithms and the human eye to identify. The algorithm compares continuous feature images frame by frame. When different levels of feature threshold conditions are met, it can automatically determine different levels of abnormal features and automatically send an alarm signal to the conveyor belt linkage device. At the same time, to improve the detection efficiency and energy saving needs of the equipment, the system automatically obtains the real-time operation signal of the conveyor belt. Its detection switch is consistent with the start-stop state, that is, detection is carried out when the conveyor is started, and detection is stopped otherwise. The user can effectively focus on the working condition of the conveyor in the running state.

[0045] 2. After the system is deployed and debugged, the alarm threshold is set in advance according to the belt running speed. Image recognition can be carried out continuously around the clock. When abnormal features are detected on the belt surface, the system automatically issues a graded alarm and records screenshots in the database for query. In this scenario, its technical feature is that the system does not require long-term training after routine debugging. It has the advantages of good image effect, high tear recognition rate, strong algorithm accuracy, and long-term stable use. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0047] Figure 1 This is a schematic diagram of the overall process provided in Example 1 of the present invention;

[0048] Figure 2 A first flow chart of the first embodiment of the present invention is provided;

[0049] Figure 3 A second flow chart of the first embodiment of the present invention;

[0050] Figure 4 A schematic diagram of a first process flow provided in the second embodiment of the present invention;

[0051] Figure 5 A schematic diagram of a first process flow provided in Embodiment 3 of the present invention;

[0052] Figure 6 This is a second flow chart provided for the third embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0054] Example 1:

[0055] See also Figure 1-Figure 3 A start-stop bidirectional linkage conveyor belt longitudinal tear machine vision detection method includes:

[0056] S1. Acquire a characteristic image of longitudinal tearing at the bottom of the conveyor belt when it is running;

[0057] S2. Perform frame-by-frame image extraction, image filtering, deformation detection, image correlation decision-making, and mask processing on the bottom longitudinal tearing condition characteristic image to generate an alarm signal and a bottom longitudinal tearing condition characteristic image after abnormality processing;

[0058] S3, obtaining the alarm signal and the characteristic image of the bottom longitudinal tearing condition after abnormal processing, and realizing system linkage;

[0059] S2 includes:

[0060] S21, image extraction includes cropping and shielding complex invalid pixels on the scene, setting the feature amplification area as the algorithm ROI, and binarizing the laser morphology within the algorithm ROI to obtain image extraction data;

[0061] S22, image filtering includes image distortion correction and image denoising to obtain steady-state feature image data;

[0062] S23, performing sudden deformation edge detection on the steady-state feature image data to obtain sudden deformation data;

[0063] S24. Perform image correlation processing on the sudden deformation data to obtain an abnormality judgment representing the risk level, and output an abnormality feature classification analysis conclusion based on the abnormality judgment.

[0064] S1 includes:

[0065] S11: After the conveyor belt is running at a uniform speed, the bottom of the conveyor belt is continuously scanned without stroboscopic light, and a real-time image of the bottom of the conveyor belt is collected immediately after the conveyor belt is connected to the material trough for unloading;

[0066] S12, performing frame-by-frame processing and continuous decision-making calculation on the bottom real-time image.

[0067] The S1 also includes:

[0068] S13. Adjust according to the on-site situation to obtain a real-time image of the bottom after correcting and amplifying the longitudinal tearing characteristics.

[0069] Specific implementation method: An industrial camera, laser, and light source system are installed below the trough at the tail end of the conveyor belt. All equipment observes from bottom to top. The center of sight plane of the industrial camera intersects with the straight line formed by the laser at the bottom of the conveyor belt to obtain a characteristic image of the longitudinal tearing at the bottom of the conveyor belt when it is running.

[0070] All equipment is installed at fixed locations. As the conveyor belt moves at a constant speed, the laser and industrial light source continuously scan the bottom of the conveyor belt without flickering. The industrial camera captures clear, real-time images of the conveyor belt bottom at a high-frequency frame rate of at least 60fps, the moment the conveyor belt is connected to the trough for unloading.

[0071] Industrial cameras capture real-time images of the conveyor belt bottom after amplification with lasers and industrial light sources, and transmit the image data to the edge computing module via a 1Gbps network. The edge computing module performs frame-by-frame processing and continuous decision-making calculations.

[0072] Industrial cameras, lasers, and industrial light sources are fine-tuned according to different on-site conditions. All fine-tuning can be remotely debugged through the edge computing module to ensure that the industrial camera captures stable and clear real-time images of the features magnified by the laser and industrial light source.

[0073] Example 2:

[0074] See also Figure 4 This embodiment provides a technical solution based on the first embodiment:

[0075] Binarization processing includes:

[0076]

[0077] Among them, maxval is the maximum value, src is the grayscale image, thresh is the threshold, and the result data dst is the binarized image.

[0078] Specific implementation method: After binarization post-processing of the laser morphology within the ROI, the continuous feature images are still accompanied by strong noise, graphic jitter and distortion, and further filtering optimization is required, so image distortion correction and image denoising operations are required.

[0079] Image distortion correction includes:

[0080] xdistorted=x(1+k1r2+k2r4+k3r6);

[0081] ydistorted=y(1+k1r2+k24+k3r6);

[0082] x and y represent pixels in different distortion directions, r represents the distortion radius, and k represents the distortion coefficient at different levels of visual center distance.

[0083] Image denoising includes:

[0084] The image extracted data is smoothed using a convolution filter and then processed by OpenCV's cv2.GaussianBlur function, including:

[0085] Smoothed image=cv2.GaussinaBlur(image,(x,y),ksize);

[0086] image represents the full frame pixels of the input image, x and y represent the filter size, and ksize represents the smoothing strength.

[0087] Specific implementation method: Image denoising generally uses a smoothing algorithm to reduce noise. It is necessary to retain important feature information while attenuating the noise in the image, and provide an image preprocessing method with steady-state data for deformation point search in subsequent processes. Smoothing is usually performed through a convolution filter, generally through OpenCV's cv2.Gauss i anBlur function.

[0088] Sudden deformation edge detection uses gradient calculation to detect the direction and degree of data change within the pixel and locate the area where the feature target in the image changes sharply. The gradient calculation is implemented using the Sobel operator of OpenCV, which specifically includes:

[0089] Sobel x=cv2.Sobel(Smoothed image,cv2.CV 64F,1,0,ksize);

[0090] Sobel y=cv2.Sobel(Smoothed image,cv2.CV 64F,1,0,ksize);

[0091] Smoothed image represents the steady-state feature image data. 1,0 and 0,1 are used to calculate the gradients in the x-direction and y-direction respectively. ksize represents the gradient calculation strength.

[0092] Image correlation processing associates consecutive frames and consecutive pixel clusters, and then sets gradient detection thresholds in the associated x and y directions.

[0093] S2 also includes:

[0094] S25. Generate a corresponding alarm signal based on the abnormal feature hierarchical analysis conclusion, perform mask processing on the bottom longitudinal tearing condition characteristic image, mark the abnormality, obtain the bottom longitudinal tearing condition characteristic image after abnormal processing, and send the bottom longitudinal tearing condition characteristic image after abnormal processing to the user terminal.

[0095] Specific implementation method: The alarm signal is generally output through a switch signal or http network protocol data. The characteristic image of the bottom longitudinal tearing after abnormal processing is automatically converted into a video stream image in H.264 format through the FFmpeg live video encoding system and output to the external platform.

[0096] Example 3:

[0097] See also Figure 5-Figure 6 This embodiment provides a technical solution based on the second embodiment:

[0098] S3 includes:

[0099] S31, realize two-way real-time communication with the main coal power management system platform;

[0100] S32, control the conveyor belt longitudinal tear machine vision detection to be consistent with the conveyor belt start and stop status;

[0101] S33. Send a prompt instruction to the user terminal according to the alarm signal, and receive an operation instruction from the user terminal.

[0102] Specific implementation method: Equipment supporting this visual inspection method was deployed on-site, specifically 17 sets of conveyor belt longitudinal tear machine vision inspection equipment, installed on 14 conveyor belt sites. The conveyor belt start and stop control signals are stored only in the SIS (Safety Instrumented System) database managed by the central control station. Based on the characteristic images, alarm signals and data of the longitudinal tear condition after abnormal processing, two-way linkage communication is carried out with the conveyor belt start and stop system;

[0103] The central server cluster's conveyor belt longitudinal tear machine vision detection aggregation module operates a centralized management platform system capable of aggregating 17 or more conveyor belt longitudinal tear machine vision detection devices. Relying on high-performance server processing chips, the system uses a built-in H.264 decoder to achieve a decoding frame rate of 60fps, consistent with its front-end devices. The database uses a MySQL relational database, which can store data for more than one year and has a complete HTTP protocol interface. It generally obtains front-end detection alarm data in milliseconds and simultaneously communicates in real time with the main coal power management system platform.

[0104] The machine vision detection of conveyor belt longitudinal tear is consistent with the start and stop status of the conveyor belt. It is mainly based on the HTTP protocol. The centralized management platform system of the conveyor belt longitudinal tear machine vision detection equipment is used as the client. It actively obtains the belt start and stop status signal from the SIS database and aligns the detection switch with its signal. That is, when the conveyor belt starts running, the machine vision detection of conveyor belt longitudinal tear is started, otherwise it is stopped.

[0105] The machine vision detection of longitudinal tears in conveyor belts can push alarm data and signals to the main coal-fired power management system platform or other management systems that need to be linked. The centralized management platform system of the machine vision detection equipment for longitudinal tears in conveyor belts serves as the client. When any device generates an alarm signal during the operation of the conveyor belt, the alarm is actively pushed to the main coal-fired power management system database through the http protocol, and the management personnel of the central control station are promptly prompted to make linkage processing. At the same time, the switch electrical signal can be output for the use of expanded access alarms.

[0106] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of such features. "Multiple" means two or more, unless otherwise specifically defined.

[0107] In the present invention, unless otherwise expressly specified or limited, terms such as "mounted," "connected," "connect," and "fixed" should be understood broadly. For example, they may refer to fixed connection, detachable connection, or integration; they may refer to direct connection or indirect connection through an intermediate medium; they may refer to internal communication between two components or interaction between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0108] In the present invention, unless otherwise expressly specified or limited, when a first feature is "above" or "below" a second feature, it may mean that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediary. Furthermore, when a first feature is "above," "above," or "above" a second feature, it may mean that the first feature is directly above or diagonally above the second feature, or simply means that the first feature is at a higher level than the second feature. When a first feature is "below," "below," or "below" a second feature, it may mean that the first feature is directly below or diagonally below the second feature, or simply means that the first feature is at a lower level than the second feature.

[0109] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0110] In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other.

[0111] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A start-stop bidirectional linkage conveyor belt longitudinal tear machine vision detection method, characterized in that: include: S1. Obtaining a characteristic image of the longitudinal tearing condition at the bottom of the conveyor belt when it is in operation; S2, performing frame-by-frame image extraction, image filtering, deformation detection, image correlation decision and mask processing on the bottom longitudinal tearing condition characteristic image to generate an alarm signal and a bottom longitudinal tearing condition characteristic image after abnormal processing; S3, obtaining the alarm signal and the characteristic image of the bottom longitudinal tearing condition after abnormal processing, and realizing system linkage; S2 includes: S21, the image extraction includes cutting and shielding complex invalid pixels on the scene, setting the feature amplification area to the algorithm RO I, and binarizing the laser morphology within the algorithm RO I to obtain image extraction data; S22, the image filtering includes image distortion correction and image denoising to obtain steady-state characteristic image data; S23, performing sudden deformation edge detection on the steady-state characteristic image data to obtain sudden deformation data; S24, performing image correlation processing on the sudden deformation data to obtain an abnormality judgment representing a risk level, and outputting an abnormality feature classification analysis conclusion according to the abnormality judgment.

2. According to the start-stop bidirectional linkage conveyor belt longitudinal tear machine vision detection method of claim 1, it is characterized in that: The S1 includes: S11. After the conveyor belt is running at a uniform speed, the bottom of the conveyor belt is continuously scanned without flickering, and a real-time image of the bottom of the conveyor belt is collected at the moment when the conveyor belt is connected to the material trough for unloading; S12, performing frame-by-frame processing and continuous decision-making calculation on the bottom real-time image.

3. According to the start-stop bidirectional linkage conveyor belt longitudinal tear machine vision detection method of claim 2, it is characterized in that: The S1 further comprises: S13. Adjust according to the on-site situation to obtain a real-time image of the bottom after correcting and amplifying the longitudinal tearing characteristics.

4. The method for detecting longitudinal tear of a conveyor belt with a start-stop bidirectional linkage machine vision according to claim 1 is characterized in that: The binarization process comprises: Among them, maxval is the maximum value, src is the grayscale image, thresh is the threshold, and the result data dst is the binarized image.

5. The method for detecting longitudinal tear of a conveyor belt with a start-stop bidirectional linkage machine vision according to claim 1 is characterized in that: The image distortion correction comprises: x distorted =x(1+k1r 2 +k2r 4 +k3r 6 ); and distorted =y(1+k1r 2 +k2r 4 +k3r 6 ); x and y represent pixels in different distortion directions, r represents the distortion radius, and k represents the distortion coefficient at different levels of visual center distance.

6. The method for detecting longitudinal tear of a conveyor belt with a start-stop bidirectional linkage machine vision according to claim 1 is characterized in that: The image denoising comprises: The image extraction data is smoothed using a convolution filter and then processed by the cv2.GaussianBlur function of OpenCV, including: Smoothed image=cv2.GaussinaBlur(image,(x,y),ksize); image represents the full frame pixels of the input image, x and y represent the filter size, and ksize represents the smoothing strength.

7. The method for detecting longitudinal tear of a conveyor belt with a start-stop bidirectional linkage machine vision according to claim 1 is characterized in that: The sudden deformation edge detection uses gradient calculation to detect the direction and degree of data change in the pixel and locate the area where the feature target in the image changes sharply. The gradient calculation is implemented using the sobel operator of OpenCV, which specifically includes: Sobel x=cv2.Sobel(Smoothed image,cv2.CV 64F,1,0,ksize); Sobel y=cv2.Sobel(Smoothed image,cv2.CV 64F,1,0,ksize); Smoothed image represents steady-state feature image data. 1,0 and 0,1 calculate the gradients in the x and y directions respectively. ksize represents the gradient calculation strength.

8. The method for detecting longitudinal tear of a conveyor belt with a start-stop bidirectional linkage machine vision according to claim 1 is characterized in that: The image correlation processing performs correlation binding on continuous frames and continuous pixel clusters, and then sets gradient detection thresholds in the associated x-direction and y-direction.

9. The method for detecting longitudinal tear of a conveyor belt with a start-stop bidirectional linkage machine vision according to claim 1 is characterized in that: The S2 further includes: S25. According to the conclusion of the abnormal feature classification analysis, a corresponding alarm signal is formed, the bottom longitudinal tearing condition characteristic image is masked, and the abnormality is marked to obtain the bottom longitudinal tearing condition characteristic image after the abnormality processing, and the bottom longitudinal tearing condition characteristic image after the abnormality processing is sent to the user terminal.

10. The start-stop bidirectionally linked conveyor belt longitudinal tear machine vision detection method according to claim 9, characterized in that: The S3 includes: S31, realize two-way real-time communication with the main coal power management system platform; S32, control the conveyor belt longitudinal tearing machine vision detection to be consistent with the conveyor belt start and stop state; S33, sending a prompt instruction to the user terminal according to the alarm signal, and receiving an operation instruction from the user terminal.

Images