Intelligent belt crack and tear detection system based on machine vision and laser coupling

Abstract

The invention discloses an intelligent belt crack and tear detection system based on machine vision and laser coupling, and relates to the technical field of belt conveying. The intelligent belt crack and tear detection system based on machine vision and laser coupling comprises a visual detection module, an image processing flow module, a laser-assisted detection module and a data processing and control module, and the visual detection module comprises double-view-angle linear array cameras (a camera 1 and a camera 2), an area array monitoring camera and an intelligent light source system; the image processing flow module comprises multi-modal feature fusion, dynamic threshold segmentation, morphological filtering and deep learning classification; the laser-assisted detection module comprises a laser array sensor; the data processing and control module comprises an edge computing unit, an upper computer system, an integrated expert system, early warning and linkage control, system automation and intelligentization, operation and maintenance cost reduction, unmanned detection, full-process automatic operation, manual inspection frequency reduction and manpower cost saving by more than 50%.

Description

Technical Field

[0001] The present invention relates to the technical field of belt conveying, and particularly relates to an intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser. Background Art

[0002] In industrial production, belt conveyors are core equipment for material transportation and are widely used in fields such as mines, metallurgy, and ports. However, during long-term operation, belts are prone to faults such as cracks and tears, and traditional detection methods have significant defects:

[0003] Low efficiency of manual inspection: Relying on manual visual inspection, it is highly subjective and cannot perform real-time monitoring, making it easy to overlook minor cracks.

[0004] Limitations of single-point detection: Traditional sensors (such as infrared and ultrasonic) can only detect local damage and cannot cover the entire surface of the belt.

[0005] High false alarm rate: Single detection methods are easily affected by environmental interferences (such as dust and light changes), leading to misjudgments or missed detections.

[0006] Serious consequences of accidents: Belt tears may cause material leakage, equipment shutdown, and even safety accidents, resulting in huge economic losses.

[0007] Therefore, there is an urgent need for an efficient, accurate, and real-time belt detection technology to meet the requirements of industrial safety production. Summary of the Invention

[0008] The purpose of the present invention is to solve at least one of the technical problems existing in the prior art, and provide an intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser, which can solve the problems of low efficiency of manual inspection and limitations of single-point detection.

[0009] To achieve the above purpose, the present invention provides the following technical solution: An intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser, including a visual detection module, an image processing flow module, a laser-assisted detection module, and a data processing and control module. The visual detection module includes dual-view linear array cameras (Camera 1 and Camera 2), a planar array monitoring camera (Camera 3), and an intelligent light source system;

[0010] The image processing flow module includes: multi-modal feature fusion, dynamic threshold segmentation, morphological filtering, and deep learning classification;

[0011] The laser-assisted detection module includes a laser array sensor;

[0012] The data processing and control module includes an edge computing unit, a host computer system, an integrated expert system, and early warning and linkage control.

[0013] Preferably, for the multi-modal feature fusion: after graying the color image, edge enhancement is performed by combining the gradient magnitude and direction features.

[0014] Dynamic threshold segmentation: An improved adaptive threshold segmentation based on histogram equalization and the Otsu algorithm is used to extract crack features under complex backgrounds.

[0015] Morphological filtering: Alternate opening and closing operations are used to eliminate noise, and crack geometric parameters (length, width, curvature) are obtained through a skeleton extraction algorithm.

[0016] Deep learning classification: The improved YOLOv5s model is introduced to achieve intelligent classification of cracks / scratches with an accuracy of 98.7%.

[0017] Preferably, for the laser array sensor: 16 groups of laser emitters and receivers are arranged equidistantly along the transverse direction under the belt to form a laser curtain.

[0018] Preferably, for the edge computing unit: It integrates a high-performance GPU to achieve real-time image processing (frame rate ≥ 25fps), and a fuzzy logic controller is built in to dynamically adjust the camera exposure parameters and light source intensity.

[0019] The host computer system: A human-computer interaction interface developed with C# is used to display the belt running status, detection results, and historical data curves in real time.

[0020] An integrated expert system is used to predict the remaining service life based on the crack propagation model.

[0021] Early warning and linkage control: A three-level alarm mechanism (yellow warning, orange warning, red shutdown) is supported to be linked with the PLC system to achieve automatic marking of the fault location and emergency braking of the belt.

[0022] Preferably, for the dual-view linear array cameras (Camera 1 and Camera 2): High-precision linear array cameras are used, which are respectively vertically installed beside the guide wheels on both sides of the belt, forming a 90° angle with the belt running direction, for collecting high-definition images of the belt side.

[0023] Preferably, for the area array monitoring camera (Camera 3): It is deployed at the gap of the idler under the belt.

[0024] Preferably, for the intelligent light source system: It includes a ring-shaped LED light source (matched with Camera 1 / 2) and a back-projection light source (matched with Camera 3), and the ambient light interference is eliminated through an adaptive dimming algorithm.

[0025] Preferably, for the light source system: The ring-shaped LED light sources (for Camera 1 and Camera 2) and the back-projection light source (for Camera 3) eliminate the ambient light interference through adaptive dimming.

[0026] Preferably, the laser detection module: Horizontally arrange 16 groups of laser emitters and receivers below the belt to form a laser curtain for monitoring the falling of slag.

[0027] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0028] 1. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser is automated and intelligent, reducing operation and maintenance costs and enabling unmanned detection: The whole process runs automatically, reducing the frequency of manual inspections and saving more than 50% of labor costs;

[0029] Adaptive environment adjustment: The intelligent light source system dynamically adjusts parameters according to the ambient light intensity to ensure stable detection accuracy within the illumination range of 0 - 10000 lux;

[0030] Predictive maintenance: Based on the crack propagation model and digital twin technology, predict the remaining service life of the belt, optimize the maintenance plan, and extend the equipment operation cycle.

[0031] 2. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser can monitor and give early warnings in real time, reducing accident risks. Dynamic response: The system processes images in real time at a frame rate of ≥25 fps, can trigger an early warning instantly when a crack appears, and discover potential tearing risks 60% earlier;

[0032] Three - level alarm mechanism: Provide clear decision - making basis for operation and maintenance personnel through yellow early warning (crack early warning), orange warning (crack expansion), and red shutdown (tear confirmation);

[0033] Linkage control: Link with the PLC system, automatically mark the fault location and apply emergency braking to avoid the expansion of accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention will be further described below with reference to the drawings and embodiments:

[0035] Figure 1 It is a schematic diagram of the equipment structure of the present invention;

[0036] Figure 2 It is a schematic diagram of the system flow of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] This part will describe the specific embodiments of the present invention in detail. The preferred embodiments of the present invention are shown in the drawings. The role of the drawings is to supplement the description of the text part of the specification, enabling people to intuitively and vividly understand each technical feature and the overall technical solution of the present invention, but it cannot be understood as a limitation of the protection scope of the present invention.

[0038] In the description of the present invention, it should be understood that for the orientation description, such as the orientation or positional relationship indicated by up, down, front, back, left, right, etc., it is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention.

[0039] In the description of the present invention, greater than, less than, exceeding, etc. are understood as not including the number itself, and above, below, within, etc. are understood as including the number itself. If the first and second are described only for the purpose of distinguishing technical features, they should not be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or the sequence relationship of the indicated technical features.

[0040] In the description of the present invention, unless otherwise clearly defined, terms such as set, install, connect, etc. should be understood in a broad sense. Those skilled in the art can reasonably determine the specific meanings of the above terms in the present invention in combination with the specific content of the technical solution.

[0041] Please refer to Figure 1-2 , the present invention provides a technical solution: an intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser. System architecture design: This system adopts a modular design and is mainly composed of a vision detection module, a laser-assisted detection module, and a data processing and control module; the system architecture is as Figure 2 shown, and each module realizes data interaction and collaborative work through an industrial bus;

[0042] An intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser includes a vision detection module, an image processing flow module, a laser-assisted detection module, and a data processing and control module;

[0043] The vision detection module includes a dual-view linear array camera (Camera 1 and Camera 2), a area array monitoring camera (Camera 3), and an intelligent light source system;

[0044] The image processing flow module includes: multi-modal feature fusion, dynamic threshold segmentation, morphological filtering, and deep learning classification;

[0045] Multi-modal feature fusion: After graying the color image, edge enhancement is performed by combining the gradient amplitude and direction features;

[0046] Dynamic threshold segmentation: An improved adaptive threshold segmentation based on histogram equalization and the Otsu algorithm is used to extract crack features under complex backgrounds;

[0047] Morphological filtering: Alternate opening and closing operations are used to eliminate noise, and crack geometric parameters (length, width, curvature) are obtained through a skeleton extraction algorithm;

[0048] Deep learning classification: Introduce the improved YOLOv5s model to achieve intelligent classification of cracks / scratches with an accuracy of 98.7%;

[0049] Laser-assisted detection module, including a laser array sensor;

[0050] Laser array sensor: 16 groups of laser emitters and receivers are arranged equidistantly along the transverse direction under the belt to form a laser curtain;

[0051] Detection principle of the laser-assisted detection module:

[0052] During normal operation, the laser beam penetrates the space under the belt to form a stable optical signal;

[0053] When the belt tears and slag drops, the laser beam is blocked, resulting in a sudden change in light intensity, triggering a tear signal;

[0054] Combined with the time difference positioning algorithm, the tear position can be accurately calculated with an error ≤ 5 cm;

[0055] Data processing and control module, including an edge computing unit, a host computer system, an integrated expert system, early warning and linkage control;

[0056] Edge computing unit: Integrate a high-performance GPU to achieve real-time image processing (frame rate ≥ 25 fps), and built-in fuzzy logic controller to dynamically adjust the camera exposure parameters and light source intensity;

[0057] Host computer system: A human-computer interaction interface developed with C# to display the belt running status, detection results, and historical data curves in real time;

[0058] Integrated expert system, predicting the remaining service life based on the crack propagation model;

[0059] Early warning and linkage control: A three-level alarm mechanism (yellow warning, orange warning, red shutdown), supporting linkage with the PLC system to achieve automatic marking of the fault location and emergency braking of the belt;

[0060] Dual-view linear array cameras (Camera 1 and Camera 2): High-precision linear array cameras are respectively vertically installed beside the guide wheels on both sides of the belt at a 90° angle to the belt running direction for collecting high-definition images of the belt side;

[0061] Area array monitoring camera (Camera 3): Deployed at the gap of the idler under the belt to capture images of the slag dropped during belt tearing;

[0062] Intelligent light source system: Includes a ring LED light source (matched with Camera 1 / 2) and a back-projection light source (matched with Camera 3), eliminating environmental light interference through an adaptive dimming algorithm;

[0063] Furthermore, the visual detection module:

[0064] Cameras 1 and 2: Vertically installed beside the guide wheels on both sides of the belt, they collect high-definition images of the left and right sides of the moving belt in real time;

[0065] Camera 3: Located in the gap between the idlers under the belt, it captures the images of the slag falling when the belt tears;

[0066] Light source system: The ring-shaped LED light source (for Cameras 1 and 2) and the back-projection light source (for Camera 3) eliminate ambient light interference through adaptive dimming;

[0067] Laser detection module: 16 groups of laser emitters and receivers are arranged horizontally under the belt to form a laser curtain for monitoring the falling of slag;

[0068] Image processing and feature extraction (for Cameras 1 and 2):

[0069] Image preprocessing:

[0070] Color image grayscale conversion: Convert the RGB image to a single-channel grayscale image to simplify the calculation;

[0071] Dynamic contrast enhancement: Improve the visibility of the crack area through histogram equalization;

[0072] Adaptive threshold segmentation: Combine the Otsu algorithm with local thresholds to segment cracks from the background;

[0073] Edge and morphological processing:

[0074] Canny edge detection: Extract the belt edge and the potential crack contour;

[0075] Dilation - erosion operation: Enlarge the crack features and remove noise interference;

[0076] Feature measurement and classification:

[0077] Skeleton extraction: Obtain the geometric parameters of the crack (length, width, curvature);

[0078] Deep learning classification: Use the improved YOLOv5 model to distinguish cracks from scratches;

[0079] Slag detection and signal fusion:

[0080] Camera 3 detection: Analyze the image under the belt to identify the falling of slag;

[0081] Laser detection: The occlusion of the laser beam by slag triggers a signal, and the tearing position is located by combining the time difference (error ≤ 5 cm);

[0082] Multi-source data fusion: Spatially and temporally synchronize and verify the results of visual crack analysis with the laser slag signal to reduce misjudgment;

[0083] Decision-making and control:

[0084] Edge computing unit: Process image data in real time (frame rate ≥ 25fps) and dynamically adjust camera parameters;

[0085] Host computer system:

[0086] Display the belt status, detection results, and historical data curves;

[0087] Predict the remaining life based on the crack propagation model;

[0088] Early warning and linkage:

[0089] Three-level alarm (yellow warning, orange warning, red shutdown);

[0090] Link with the PLC system to automatically mark the fault location and apply emergency braking;

[0091] Figure 1 Corresponding description of the operation process:

[0092] Visual detection area: Cameras 1 and 2 cooperate with the light source to collect images of the belt side, and after being processed by the edge computing unit, they are transmitted to the host computer;

[0093] Laser detection area: The laser array and camera 3 form a cross-detection under the belt, and the slag trigger signal is transmitted to the PLC through the RS485 bus;

[0094] Data interaction: The edge computing unit communicates with the host computer through industrial Ethernet to achieve real-time sharing of detection results and issuance of control instructions;

[0095] Furthermore, the multi-modal fusion detection of this system improves accuracy and reliability. Visual detection: Use dual-view cameras (Camera 1 / 2) to collect images of the belt side, and combine Canny edge detection, morphological filtering, and deep learning algorithms to accurately identify millimeter-level cracks and distinguish cracks from scratches;

[0096] Laser-assisted verification: The laser amplifier and camera 3 cooperate to detect falling slag, and through spatio-temporal synchronization verification, reduce the misjudgment rate of tear detection to 0.3 times per 10,000 hours;

[0097] Data fusion: Fuse visual image features and laser signals to achieve crack propagation trend analysis, with the accuracy rate increased to 98.7%;

[0098] Furthermore, the system can monitor and give early warnings in real time, reducing accident risks and making dynamic responses: The system processes images in real time at a frame rate of ≥25fps, can trigger an early warning the moment a crack appears, and detect potential tearing risks 60% earlier.

[0099] Three-level alarm mechanism: Through yellow early warning (crack early warning), orange warning (crack expansion), and red shutdown (tearing confirmation), it provides clear decision-making basis for operation and maintenance personnel.

[0100] Linkage control: Link with the PLC system, automatically mark the fault location and apply emergency braking to prevent the accident from expanding.

[0101] Furthermore, the system is automated and intelligent, reducing operation and maintenance costs. Unattended detection: The whole process runs automatically, reducing the frequency of manual inspections and saving more than 50% of the labor cost.

[0102] Adaptive environment adjustment: The intelligent light source system dynamically adjusts parameters according to the ambient light intensity to ensure stable detection accuracy within the illumination range of 0 - 10000 lux.

[0103] Predictive maintenance: Based on the crack propagation model and digital twin technology, predict the remaining service life of the belt, optimize the maintenance plan, and extend the equipment operation cycle.

[0104] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the knowledge scope of those of ordinary skill in the art in the said technical field, various changes can be made without departing from the purpose of the present invention.

Claims

1. An intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser, comprising a vision detection module, an image processing flow module, a laser-assisted detection module, and a data processing and control module, characterized in that: The visual detection module includes dual-view linear array cameras (Camera 1 and Camera 2), a matrix monitoring camera (Camera 3), and an intelligent light source system; The image processing flow module includes: multi-modal feature fusion, dynamic threshold segmentation, morphological filtering, and deep learning classification; The laser-assisted detection module includes a laser array sensor; The data processing and control module includes an edge computing unit, a host computer system, an integrated expert system, warning and linkage control.

2. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser according to claim 1, wherein: The multi-modal feature fusion: After graying the color image, edge enhancement is performed by combining gradient magnitude and direction features; Dynamic threshold segmentation: An improved adaptive threshold segmentation based on histogram equalization and Otsu algorithm is used to extract crack features under complex backgrounds; Morphological filtering: Alternate opening and closing operations are used to eliminate noise, and crack geometric parameters (length, width, curvature) are obtained through a skeleton extraction algorithm; Deep learning classification: The improved YOLOv5s model is introduced to achieve intelligent classification of cracks / scratches, with an accuracy of 98.7%; 3. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser according to claim 1, characterized in that: The laser array sensor: 16 groups of laser emitters and receivers are arranged equidistantly along the transverse direction below the belt to form a laser curtain; 4. The intelligent belt crack and tear detection system based on the coupling of machine vision and laser according to claim 1, wherein: The edge computing unit: Integrates a high-performance GPU to achieve real-time image processing (frame rate ≥ 25fps), and a fuzzy logic controller is built-in to dynamically adjust the camera exposure parameters and light source intensity; The host computer system: A human-computer interaction interface developed in C# is used to display the belt running status, detection results, and historical data curves in real time; The integrated expert system predicts the remaining service life based on the crack propagation model; Warning and linkage control: A three-level alarm mechanism (yellow warning, orange warning, red shutdown), supports linkage with the PLC system, and realizes automatic marking of the fault location and emergency braking of the belt.

5. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser according to claim 1, wherein: The dual-view linear array cameras (Camera 1 and Camera 2): High-precision linear array cameras are used, which are respectively vertically installed beside the guide wheels on both sides of the belt, forming a 90° angle with the belt running direction, and are used to collect high-definition images of the belt side; 6. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser according to claim 1, characterized in that: The matrix monitoring camera (Camera 3): Is deployed at the gap of the idler below the belt; 7. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser according to claim 1, characterized in that: The intelligent light source system: Includes a ring-shaped LED light source (matched with Camera 1 / 2) and a back-projection light source (matched with Camera 3), and eliminates ambient light interference through an adaptive dimming algorithm; 8. The intelligent detection system for belt cracks and tears based on the coupling of machine vision and laser according to claim 1, characterized in that: The light source system: The ring-shaped LED light sources (Camera 1 and Camera 2) and the back-projection light source (Camera 3) eliminate ambient light interference through adaptive dimming; 9. The intelligent belt crack and tear detection system based on the coupling of machine vision and laser according to claim 1, characterized in that: The laser detection module: 16 groups of laser emitters and receivers are arranged horizontally below the belt to form a laser curtain to monitor the falling of slag.

Images