Coal conveying belt torn edge tearing detection method and system based on edge recognition

Abstract

The invention discloses a coal conveying belt torn edge tearing detection method and system based on edge recognition, and relates to the technical field of belt torn edge tearing detection, and the method comprises the steps: obtaining a coal conveying belt on-site image, detecting a coal conveying belt through a deep learning model, and extracting an ROI region; carrying out median filtering denoising and histogram equalization enhancement on the ROI region, and extracting crack features by adopting Canny edge detection; whether the belt is torn or not is judged based on the extracted edge features, if yes, warning information is pushed to a monitoring platform, and if not, detection continues. According to the coal conveying belt torn edge tearing detection method based on edge recognition, the crack characteristics are extracted by combining Canny edge detection with morphological analysis, the tearing degree is judged by setting the threshold value, the crack detection accuracy is improved, and the false detection rate is reduced. The method has better effects in the aspects of real-time performance, detection precision and adaptability to complex working conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of belt edge tear detection, and in particular to a method and system for detecting edge tear of a coal conveyor belt based on edge recognition. Background Art

[0002] Coal-fired power plants often use coal conveyor belts as key equipment for long-distance coal transportation. During the production and transportation process, belt failures, such as belt tearing and belt deviation, are common due to factors such as long-term operation, continuous belt stress, weather conditions, and coal quality.

[0003] Belt conveyors used in production generally operate at high speeds. If a belt tear occurs and the machine is not shut down immediately or field personnel are notified to identify and eliminate the source of the tear, a through-tear can occur. If the entire belt tears, it can spill material, damage equipment like reducers and motors, and cause factory shutdowns, significantly impacting power safety. In severe cases, it can even damage the machine frame, threatening the safety of field personnel and resulting in significant economic losses.

[0004] Currently, belt anomalies, such as belt tears, are mostly detected through manual inspections or contact-based detection equipment (tear ball switches). However, in actual production, small belt tears and perforations are difficult to detect at an early stage. Furthermore, limited manpower and the triggering limitations of ball-type tear switches make it difficult to accurately monitor the belt's operating status in real time. Once a small tear occurs in a rubber belt, it only takes a few turns for the belt to fully tear. Therefore, by the time the belt anomaly is discovered, significant equipment losses have already occurred and are irreversible. Remediation can take one to three days. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is: the existing coal conveyor belt tear detection method relies on manual inspection, has poor real-time performance, low detection accuracy, and difficulty in identifying early tears. Traditional contact detection equipment has large triggering limitations and cannot fully cover the entire belt. The problem also arises of how to achieve high-precision, real-time, non-contact monitoring of coal conveyor belt tears based on computer vision and deep learning technology.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for detecting torn edges of coal conveyor belts based on edge recognition, comprising obtaining on-site images of the coal conveyor belts, detecting the coal conveyor belts using a deep learning model and extracting ROI areas.

[0008] The ROI area is subjected to median filtering denoising and histogram equalization enhancement, and Canny edge detection is used to extract crack features.

[0009] Based on the extracted edge features, it is determined whether the belt is torn. If it is torn, an alarm message is pushed to the monitoring platform; otherwise, the detection continues.

[0010] As a preferred solution of the method for detecting torn edges and tears of a coal conveyor belt based on edge recognition described in the present invention, the acquisition of on-site images of the coal conveyor belt includes acquiring a video stream of the coal conveyor belt through a high-definition camera and transmitting the video stream to an image algorithm server.

[0011] As a preferred solution of the method for detecting torn edges and tears of coal conveyor belts based on edge recognition described in the present invention, the method of detecting the coal conveyor belt and extracting the ROI area using a deep learning model includes using a YOLO target detection network to detect the coal conveyor belt, and obtaining the belt ROI area through bounding box regression, while using non-maximum suppression to remove duplicate detection frames.

[0012] As a preferred solution of the edge recognition-based coal conveyor belt tear detection method described in the present invention, the median filtering denoising and histogram equalization enhancement of the ROI area include adopting adaptive histogram equalization, processing the image by block, and setting contrast limits.

[0013] As a preferred solution of the edge recognition-based coal conveyor belt tear detection method described in the present invention, the method of extracting crack features using Canny edge detection includes calculating the connected area, edge gradient amplitude and edge shape factor of the crack to remove misjudgments caused by stains, lighting changes or background interference.

[0014] As a preferred solution of the method for detecting torn edges and tears of a coal conveyor belt based on edge recognition described in the present invention, the method of judging whether the belt is torn based on the extracted edge features includes judging whether the belt is torn based on set threshold conditions, including crack length, width and area. When the crack length is greater than the preset threshold and the area of the connected area exceeds the set value, it is judged that the belt is torn.

[0015] As a preferred solution of the edge recognition-based coal conveyor belt tear detection method described in the present invention, the pushing of alarm information to the monitoring platform includes adopting the MQTT or WebSocket protocol to transmit the tear detection results to the monitoring platform, and displaying the specific location, size information, occurrence time and historical trend analysis of the tear on the platform.

[0016] Another object of the present invention is to provide a coal conveyor belt edge tear detection system based on edge recognition, which can extract belt crack characteristics by using Canny edge detection, and calculate the length, width and connectivity of the cracks in combination with morphological analysis, and set a threshold to judge the degree of tearing, thereby solving the problems of traditional detection methods that are insufficient in recognizing small cracks, have a high false detection rate, and are difficult to distinguish between stains and real cracks.

[0017] As a preferred embodiment of the edge recognition-based coal conveyor belt tear detection system described in the present invention, it includes a path planning module, a PID control module, and a smoothing processing module. The image extraction module is used to obtain on-site images of the coal conveyor belt, detect the coal conveyor belt using a deep learning model, and extract the ROI region. The feature extraction module is used to perform median filtering denoising and histogram equalization enhancement on the ROI region, and use Canny edge detection to extract crack features. The tear analysis module is used to determine whether the belt is torn based on the extracted edge features. If so, an alarm message is sent to the monitoring platform; otherwise, detection continues.

[0018] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a method for detecting torn edges and tears of a coal conveyor belt based on edge recognition.

[0019] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for detecting edge and tearing of a coal conveyor belt based on edge recognition.

[0020] The beneficial effects of the present invention are as follows: The proposed method for detecting torn edges and tears on coal conveyor belts based on edge recognition uses Canny edge detection combined with morphological analysis to extract crack features and sets a threshold to determine the degree of tearing. This improves crack detection accuracy and reduces false positives. Based on a deep learning object detection model (YOLO), the method locates the coal conveyor belt area and extracts a ROI for further analysis. The method can automatically adapt to varying lighting and background complexity, improving detection robustness. This method achieves superior real-time performance, detection accuracy, and adaptability to complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1This is an overall flow chart of a method for detecting torn edges of a coal conveyor belt based on edge recognition, provided in the first embodiment of the present invention. DETAILED DESCRIPTION

[0023] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0024] Example 1, with reference to Figure 1 , as one embodiment of the present invention, provides a method for detecting torn edges of a coal conveyor belt based on edge recognition, comprising:

[0025] S1: Obtain on-site images of the coal conveyor belt, use the deep learning model to detect the coal conveyor belt and extract the ROI area.

[0026] Furthermore, obtaining on-site images of the coal conveyor belt includes obtaining a video stream of the coal conveyor belt through a high-definition camera and transmitting the video stream to an image algorithm server.

[0027] High-definition cameras (e.g. 2 megapixels or above) are installed along the coal conveyor belt to continuously capture video streams or regularly capture images to ensure that the shooting angle can cover the belt surface for subsequent processing.

[0028] The collected video stream or image is transmitted to the image algorithm server via industrial Ethernet or wireless transmission (such as 5G / LoRa) for subsequent processing.

[0029] It should be noted that the use of deep learning models to detect the coal conveyor belt and extract the ROI area includes using the YOLO target detection network to detect the coal conveyor belt, and obtaining the belt ROI area through bounding box regression, while using non-maximum suppression to remove duplicate detection boxes.

[0030] The YOLO V5 / V8 deep learning object detection model was used. Detection accuracy was improved using a training dataset (including images of coal conveyor belts with varying lighting, backgrounds, and contamination). Transfer learning was used to fine-tune the model to adapt to different coal conveyor environments.

[0031] The belt's ROI region is determined using the YOLO model's bounding box regression. The ROI region is automatically cropped to reduce background interference and improve computational efficiency for subsequent processing. Overlapping detection frames are addressed, and redundant frames are removed to ensure the uniqueness of the ROI region.

[0032] Set the IoU (Intersection over Union) threshold to improve the model's ability to accurately locate the belt area.

[0033] S2: Perform median filtering to denoise the ROI area, perform histogram equalization enhancement, and use Canny edge detection to extract crack features.

[0034] Furthermore, the ROI region is subjected to median filtering denoising and histogram equalization enhancement including the use of adaptive histogram equalization, the image is processed by block division, and contrast limits are set.

[0035] The belt surface may generate noise due to dust, coal powder, oil, etc., which may affect the accuracy of crack detection.

[0036] Traditional Gaussian filtering will blur the crack edges, so median filtering is selected to remove noise while retaining edge features.

[0037] Adaptive median filtering using a 3×3 or 5×5 window:

[0038] The window size is dynamically adjusted according to the noise density to ensure denoising without losing crack details.

[0039] Calculate the signal-to-noise ratio (SNR) to ensure that the quality of the filtered image meets the detection requirements.

[0040] It should be noted that the use of Canny edge detection to extract crack features includes calculating the connected area of the crack, the edge gradient amplitude, and the edge shape factor to remove misjudgments caused by stains, lighting changes, or background interference.

[0041] The surface of the coal conveyor belt has uneven reflections, and some cracks may be difficult to identify due to insufficient brightness, so histogram equalization is performed to enhance the contrast.

[0042] CLAHE (Adaptive Histogram Equalization): Uses a local enhancement strategy to avoid over-enhancement caused by global histogram equalization. Sets contrast limit parameters to avoid images being too bright or too dark.

[0043] The local contrast index (LCI) is calculated to ensure that the contrast between the crack and the background is maximized after enhancement, thereby improving the accuracy of subsequent detection.

[0044] S3: Determine whether the belt is torn based on the extracted edge features. If so, send an alarm to the monitoring platform; otherwise, continue detection.

[0045] Furthermore, judging whether the belt is torn based on the extracted edge features includes judging whether the belt is torn based on set threshold conditions, including crack length, width and area. When the crack length is greater than the preset threshold and the area of the connected area exceeds the set value, it is judged that the belt is torn.

[0046] Calculate the crack length, width, and area based on the edge detection results:

[0047] Crack length > 50 pixels (adjustable).

[0048] The area of the crack connected region > the set threshold.

[0049] Calculate fracture connectivity, excluding isolated noise points:

[0050] The Connected Component Analysis (CCA) method is used to determine whether the crack is a continuous tear.

[0051] Combined with Fourier transform, the frequency characteristics of the crack edge are analyzed to eliminate the interference of stains.

[0052] Use CNN or ResNet to learn crack features and improve crack detection accuracy.

[0053] Combined with historical data, the crack expansion trend is analyzed.

[0054] It should be noted that pushing alarm information to the monitoring platform includes using the MQTT or WebSocket protocol to transmit the tear detection results to the monitoring platform, and displaying the specific location, size information, occurrence time and historical trend analysis of the tear on the platform.

[0055] Embodiment 2 is the second embodiment of the present invention, which differs from the first two embodiments in that:

[0056] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0057] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0058] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0059] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0060] Example 3 is the third embodiment of the present invention. This embodiment provides a system for detecting torn edges of coal conveyor belts based on edge recognition, including a path planning module, a PID control module, and a smoothing processing module.

[0061] The image extraction module captures on-site images of the coal conveyor belt, uses a deep learning model to detect the belt, and extracts the ROI region. The feature extraction module performs median filtering to remove noise, histogram equalization to enhance the ROI region, and uses Canny edge detection to extract crack features. The tear analysis module determines whether the belt is torn based on the extracted edge features. If so, an alarm is sent to the monitoring platform; otherwise, detection continues.

[0062] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be covered by the scope of the claims of the present invention.

Claims

1. A method for detecting torn edges of coal conveyor belts based on edge recognition, characterized in that: include: Obtain on-site images of the coal conveyor belt, use the deep learning model to detect the coal conveyor belt and extract the ROI area; The ROI area is subjected to median filtering for denoising and histogram equalization for enhancement, and Canny edge detection is used to extract crack features; Based on the extracted edge features, it is determined whether the belt is torn. If it is torn, an alarm message is pushed to the monitoring platform; otherwise, the detection continues.

2. The method for detecting torn edges of a coal conveyor belt based on edge recognition according to claim 1, characterized in that: The obtaining of on-site images of the coal conveyor belt includes obtaining a video stream of the coal conveyor belt through a high-definition camera and transmitting the video stream to an image algorithm server.

3. The method for detecting torn edges of a coal conveyor belt based on edge recognition according to claim 2, wherein: The use of a deep learning model to detect the coal conveyor belt and extract the ROI area includes using a YOLO target detection network to detect the coal conveyor belt, obtaining the belt ROI area through bounding box regression, and using non-maximum suppression to remove duplicate detection frames.

4. The method for detecting torn edges of a coal conveyor belt based on edge recognition according to claim 3, wherein: The median filtering denoising and histogram equalization enhancement of the ROI region includes adopting adaptive histogram equalization, processing the image by dividing it into blocks, and setting contrast limits.

5. The method for detecting edge tearing of a coal conveyor belt based on edge recognition according to claim 4, characterized in that: The method of extracting crack features by using Canny edge detection includes calculating the connected area, edge gradient amplitude, and edge shape factor of the crack to remove misjudgments caused by stains, illumination changes, or background interference.

6. The method for detecting torn edges of a coal conveyor belt based on edge recognition according to claim 5, characterized in that: The method of judging whether the belt is torn based on the extracted edge features includes setting threshold conditions, including crack length, width and area. When the crack length is greater than the preset threshold and the area of the connected area exceeds the set value, it is judged that the belt is torn.

7. The method for detecting edge tearing of a coal conveyor belt based on edge recognition according to claim 6, characterized in that: Pushing the alarm information to the monitoring platform includes using the MQTT or WebSocket protocol to transmit the tear detection results to the monitoring platform, and displaying the specific location, size information, occurrence time and historical trend analysis of the tear on the platform.

8. A system using the method for detecting torn edges of a coal conveyor belt based on edge recognition according to any one of claims 1 to 7, characterized in that: Including path planning module, PID control module, and smoothing processing module; The image extraction module is used to obtain on-site images of the coal conveyor belt, detect the coal conveyor belt using a deep learning model and extract the ROI area; The feature extraction module is used to perform median filtering denoising and histogram equalization enhancement on the ROI area, and use Canny edge detection to extract crack features; The tear analysis module is used to determine whether the belt is torn based on the extracted edge features. If it is torn, an alarm message is pushed to the monitoring platform; otherwise, the detection continues.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for detecting torn edges and tears of a coal conveyor belt based on edge recognition according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for detecting torn edges and tears of a coal conveyor belt based on edge recognition according to any one of claims 1 to 7 are implemented.

Images