Coal conveying belt offset detection method and system

Abstract

The invention provides a coal conveying belt offset detection method and system, and the method comprises the steps: carrying out the image processing of the obtained operation state information of a coal conveying belt through a target detection algorithm, and obtaining the visual offset information of the coal conveying belt; physical offset information of the coal conveying belt is obtained according to the obtained sensor measurement values collected by the laser distance measuring sensors at the symmetrical positions of the two sides of the coal conveying belt; performing fusion analysis according to the visual offset information and the physical offset information to obtain offset state information of the coal conveying belt; according to the offset state information of the coal conveying belt, offset correction is conducted on the coal conveying belt through a multi-source data fusion algorithm; according to the invention, by combining two non-contact detection means of computer vision detection and laser ranging, advantage complementation can be realized; therefore, by means of the method, the false alarm rate and the omission ratio in the coal conveying belt deviation detection process can be reduced, and more reliable coal conveying belt deviation detection can be achieved.

Description

Technical Field

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

[0002] Currently, in industries such as coal, electricity, and ports, coal conveyor belts are key equipment in material conveying systems, and their stable operation is of great significance for ensuring the continuity and safety of the entire production system. However, during long-term operation, coal conveyor belts often deviate (i.e., "deviate") due to various factors such as uneven material distribution, roller damage, loading shock, changes in environmental conditions (such as temperature and humidity), and installation errors. This deviation not only causes coal and other materials to spill, resulting in material loss and environmental pollution, increasing the workload for cleaning and operation and maintenance costs, but may also cause belt scratches and tears, and damage to key components such as rollers and rollers, thereby shortening equipment life and even causing serious failures such as belt jamming and breakage, bringing the risk of equipment downtime and production interruption. In extreme cases, it may also cause safety accidents such as fire.

[0003] To address this issue, existing belt deviation detection technologies, while diverse, generally suffer from low detection accuracy, high false alarm rates, and poor robustness. This is particularly true in complex on-site environments, with variable lighting conditions, severe interference from dust or water vapor, obstruction of the detection target, or unstable sensor performance. False alarms or missed detections are more likely to occur, making it impossible to accurately determine whether the belt has deviated and its severity. For example, traditional manual inspections rely on operator visual judgment, which is time-consuming and prone to missed detections. While mechanical contact methods can trigger alarms, they suffer from physical wear, jamming, and delayed response, and it is difficult to quantify the degree of deviation. Electromagnetic induction methods are susceptible to electromagnetic interference or on-site dust, leading to inaccuracy. Laser ranging methods, while highly accurate, are costly and susceptible to interference from water mist or dust. Pure visual methods, while offering the advantages of non-contact and quantifiable detection, struggle to guarantee detection accuracy in situations such as obscured pulleys, drastic changes in lighting conditions, or contaminated and reflective belt surfaces. Summary of the Invention

[0004] In order to solve the problem that the existing technology is prone to false alarms or missed alarms when detecting the deviation of a coal conveyor belt under certain working conditions (harsh environment, feature occlusion, sensor failure), resulting in the inability to timely detect and deal with the deviation, the present invention proposes a coal conveyor belt deviation detection method, comprising:

[0005] Performing image processing on the acquired operating status information of the coal conveyor belt using a target detection algorithm to obtain visual offset information of the coal conveyor belt;

[0006] Obtaining physical offset information of the coal conveyor belt based on sensor measurement values ​​collected by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt;

[0007] Performing a fusion analysis based on the visual offset information and the physical offset information to obtain offset status information of the coal conveyor belt;

[0008] According to the offset state information of the coal conveyor belt, a multi-source data fusion algorithm is used to correct the offset of the coal conveyor belt.

[0009] Optionally, performing image processing on the acquired operating status information of the coal conveyor belt using a target detection algorithm to obtain visual offset information of the coal conveyor belt includes:

[0010] Based on the acquired operating status information of the coal conveyor belt, the improved YOLOv8 target detection algorithm is used to obtain pulley information of the coal conveyor belt; wherein the pulley information includes one or more of the following: bounding box coordinate information, category information, information about the number of pulleys on the left side, and information about the number of pulleys on the right side;

[0011] According to the pulley information of the coal conveyor belt, the difference between the number of pulleys on the left side and the number of pulleys on the right side of the coal conveyor belt is calculated to obtain the difference between the pulleys on both sides of the coal conveyor belt;

[0012] Determining visual offset information of the coal conveyor belt according to a difference between pulleys on both sides of the coal conveyor belt;

[0013] Among them, the backbone network of the improved YOLOv8 target detection algorithm is EfficientViT;

[0014] The visual offset information includes: a visual offset direction and a visual offset degree.

[0015] Optionally, determining the visual offset information of the coal conveyor belt according to the difference between pulleys on both sides of the coal conveyor belt includes:

[0016] If the difference between the pulleys on both sides of the coal conveyor belt is greater than a set first threshold, it is determined that the visual offset direction of the coal conveyor belt is right;

[0017] If the difference between the pulleys on both sides of the coal conveyor belt is less than a set second threshold, it is determined that the visual offset direction of the coal conveyor belt is left;

[0018] If the wheel slip difference on both sides of the coal conveyor belt is within a set first threshold range, it is determined that the coal conveyor belt has no visual deviation;

[0019] If the wheel slip difference on both sides of the coal conveyor belt is within a set second threshold range, it is determined that the visual deviation of the coal conveyor belt is slight;

[0020] If the wheel slip difference on both sides of the coal conveyor belt is within a set third threshold range, it is determined that the visual deviation degree of the coal conveyor belt is moderate;

[0021] If the wheel slip difference on both sides of the coal conveyor belt is within a set fourth threshold range, it is determined that the visual deviation degree of the coal conveyor belt is serious.

[0022] Optionally, obtaining the physical offset information of the coal conveyor belt based on the sensor measurement values ​​collected by the laser ranging sensors at the symmetrical positions on both sides of the coal conveyor belt includes:

[0023] performing difference calculation based on sensor measurement values ​​collected by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt to obtain a laser ranging difference;

[0024] If the laser ranging difference is less than a set distance threshold, it is determined that the coal conveyor belt has not physically deviated;

[0025] If the laser ranging difference is greater than or equal to the distance threshold, it is determined that the coal conveyor belt has physically deviated, and the deflection direction of the coal conveyor belt is used as the physical deflection information of the coal conveyor belt.

[0026] Optionally, the performing a fusion analysis based on the visual offset information and the physical offset information to obtain the offset state information of the coal conveyor belt includes:

[0027] When there is no visual deviation of the coal conveyor belt and the coal conveyor belt is physically deviated, a secondary verification mechanism is triggered to obtain a secondary deviation detection result of the coal conveyor belt, and the secondary deviation detection result is used as the deviation status information of the coal conveyor belt;

[0028] When the visual deviation degree of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt is physically deviated, it is determined that the coal conveyor belt is actually deviated, and the visual deviation degree and physical deviation information of the coal conveyor belt are used as the deviation state information of the coal conveyor belt;

[0029] When the visual deviation of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt has not physically deviated, it is determined that the coal conveyor belt has actually deviated, and the physical deviation information of the coal conveyor belt is used as the deviation status information of the coal conveyor belt.

[0030] Optionally, when there is no visual deviation of the coal conveyor belt and the coal conveyor belt is physically offset, triggering a secondary verification mechanism to obtain a secondary offset detection result of the coal conveyor belt includes:

[0031] When the coal conveyor belt has no visual deviation and the coal conveyor belt has physical deviation, retrieving the operation status information of the coal conveyor belt in different time frames;

[0032] According to the operating status information of the coal conveyor belt in different time frames, the improved YOLOv8 target detection algorithm is used to determine the secondary visual offset information of the coal conveyor belt, and the secondary visual offset information is used as the secondary offset detection result.

[0033] Optionally, the offset correction of the coal conveyor belt using a multi-source data fusion algorithm according to the offset state information of the coal conveyor belt includes:

[0034] According to the offset state information of the coal conveyor belt, a Kalman filter algorithm is used to predict the offset of the coal conveyor belt to obtain the offset prediction information of the coal conveyor belt;

[0035] Based on the deviation prediction information and the deviation status information of the coal conveyor belt, the deviation of the coal conveyor belt is corrected.

[0036] Based on the same inventive concept, the present invention also provides a coal conveyor belt deviation detection system, comprising:

[0037] A visual detection module is used to perform image processing on the acquired operating status information of the coal conveyor belt using a target detection algorithm to obtain visual offset information of the coal conveyor belt;

[0038] a physical detection module, configured to obtain physical offset information of the coal conveyor belt based on sensor measurement values ​​acquired by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt;

[0039] a data fusion module, configured to perform fusion analysis based on the visual offset information and the physical offset information to obtain offset status information of the coal conveyor belt;

[0040] The offset correction module is used to correct the offset of the coal conveyor belt by using a multi-source data fusion algorithm according to the offset status information of the coal conveyor belt.

[0041] Optionally, the visual inspection module includes:

[0042] A pulley detection submodule is configured to obtain pulley information of the coal conveyor belt based on the acquired operating status information of the coal conveyor belt using an improved YOLOv8 object detection algorithm; wherein the pulley information includes one or more of the following: bounding box coordinate information, category information, information about the number of pulleys on the left side, and information about the number of pulleys on the right side;

[0043] A pulley difference calculation submodule is used to calculate the difference between the number of pulleys on the left side and the number of pulleys on the right side of the coal conveyor belt according to the pulley information of the coal conveyor belt, so as to obtain the difference between the pulleys on both sides of the coal conveyor belt;

[0044] A visual offset submodule, configured to determine visual offset information of the coal conveyor belt according to a difference between pulleys on both sides of the coal conveyor belt;

[0045] Among them, the backbone network of the improved YOLOv8 target detection algorithm is EfficientViT;

[0046] The visual offset information includes: a visual offset direction and a visual offset degree.

[0047] Optionally, the visual offset submodule includes:

[0048] a first direction deviation unit, configured to determine that the visual deviation direction of the coal conveyor belt is right when the difference between the pulleys on both sides of the coal conveyor belt is greater than a set first threshold;

[0049] a second direction deviation unit, configured to determine that the visual deviation direction of the coal conveyor belt is left when the difference between the pulleys on both sides of the coal conveyor belt is less than a set second threshold;

[0050] a no-deviation output unit, configured to determine that the coal conveyor belt has no visual deviation when the wheel slip difference between the two sides of the coal conveyor belt is within a set first threshold range;

[0051] a first degree output unit, configured to determine that the visual deviation degree of the coal conveyor belt is slight when the wheel slip difference between the two sides of the coal conveyor belt is within a set second threshold range;

[0052] a second degree output unit, configured to determine that the visual deviation degree of the coal conveyor belt is moderate when the wheel slip difference on both sides of the coal conveyor belt is within a set third threshold range;

[0053] The third degree output unit is configured to determine that the visual deviation degree of the coal conveyor belt is severe when the wheel slip difference on both sides of the coal conveyor belt is within a set fourth threshold range.

[0054] Optionally, the physical detection module includes:

[0055] a laser difference calculation submodule, configured to perform difference calculation based on sensor measurement values ​​acquired by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt to obtain a laser ranging difference;

[0056] A physical no-offset submodule, configured to determine that no physical offset has occurred on the coal conveyor belt when the laser ranging difference is less than a set distance threshold;

[0057] The physical offset submodule is used to determine that the coal conveyor belt has physically offset when the laser ranging difference is greater than or equal to the distance threshold, and use the offset direction of the coal conveyor belt as the physical offset information of the coal conveyor belt.

[0058] Optionally, the data fusion module includes:

[0059] A secondary verification submodule is configured to trigger a secondary verification mechanism when there is no visual deviation of the coal conveyor belt and the coal conveyor belt is physically offset, obtain a secondary offset detection result of the coal conveyor belt, and use the secondary offset detection result as the offset status information of the coal conveyor belt;

[0060] an offset output submodule, configured to determine that the coal conveyor belt has actually offset when the visual offset degree of the coal conveyor belt is slight, moderate, or severe and the coal conveyor belt has physically offset, and to use the visual offset degree and physical offset information of the coal conveyor belt as offset status information of the coal conveyor belt;

[0061] The fuzzy judgment submodule is used to determine that the coal conveyor belt has actually shifted when the visual shift degree of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt has not physically shifted, and to use the physical shift information of the coal conveyor belt as the shift status information of the coal conveyor belt.

[0062] Optionally, the secondary syndrome module includes:

[0063] a frame information arrangement unit, configured to retrieve the operating status information of the coal conveyor belt in different time frames when the coal conveyor belt has no visual offset and the coal conveyor belt has a physical offset;

[0064] The secondary visual detection unit is used to determine the secondary visual offset information of the coal conveyor belt according to the operating status information of the coal conveyor belt in different time frames using an improved YOLOv8 target detection algorithm, and use the secondary visual offset information as the secondary offset detection result.

[0065] Optionally, the offset correction module includes:

[0066] A deviation prediction submodule is used to predict the deviation of the coal conveyor belt using a Kalman filter algorithm according to the deviation state information of the coal conveyor belt, so as to obtain the deviation prediction information of the coal conveyor belt;

[0067] The dynamic correction submodule is used to correct the offset of the coal conveyor belt based on the offset prediction information and the offset state information of the coal conveyor belt.

[0068] In another aspect, the present invention further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0069] The memory is used to store one or more programs;

[0070] When the one or more programs are executed by the at least one processor, a method for detecting deviation of a coal conveyor belt as described above is implemented.

[0071] On the other hand, the present invention further provides a computer-readable storage medium having an execution program stored thereon. When the execution program is executed, the above-mentioned method for detecting the deviation of a coal conveyor belt is implemented.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] The present invention provides a coal conveyor belt offset detection method and system, comprising: performing image processing on the acquired running status information of the coal conveyor belt using a target detection algorithm to obtain visual offset information of the coal conveyor belt; obtaining physical offset information of the coal conveyor belt based on sensor measurement values ​​collected by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt; performing fusion analysis based on the visual offset information and the physical offset information to obtain offset status information of the coal conveyor belt; and performing offset correction on the coal conveyor belt based on the offset status information of the coal conveyor belt using a multi-source data fusion algorithm. The present invention combines computer vision detection and laser ranging, two non-contact detection methods, to achieve complementary advantages. When visual detection is limited, laser ranging provides robust physical compensation, and when laser ranging is limited, visual detection provides visual compensation. Therefore, the method of the present invention can reduce false alarm rate and missed detection rate, thereby achieving more reliable coal conveyor belt offset detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 A schematic flow chart of a method for detecting deviation of a coal conveyor belt provided by the present invention;

[0075] Figure 2 A schematic diagram of the overall implementation process of a coal conveyor belt deviation detection method provided by the present invention;

[0076] Figure 3 A schematic diagram of the structure of a coal conveyor belt deviation detection system provided by the present invention;

[0077] Figure 4 This is a structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0078] The present invention provides a method, system, device and medium for detecting deviation of a coal conveyor belt. The specific embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0079] Example 1:

[0080] The present invention provides a method for detecting deviation of a coal conveyor belt, the flow diagram of which is as follows: Figure 1 As shown, including:

[0081] Step 1: Using a target detection algorithm to perform image processing on the acquired operating status information of the coal conveyor belt to obtain visual offset information of the coal conveyor belt;

[0082] Step 2: obtaining physical offset information of the coal conveyor belt based on sensor measurement values ​​collected by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt;

[0083] Step 3: performing a fusion analysis based on the visual offset information and the physical offset information to obtain offset status information of the coal conveyor belt;

[0084] Step 4: Based on the offset status information of the coal conveyor belt, a multi-source data fusion algorithm is used to correct the offset of the coal conveyor belt.

[0085] In one implementation, the process of performing image processing on the acquired operating status information of the coal conveyor belt using a target detection algorithm in step 1 to obtain visual offset information of the coal conveyor belt may include:

[0086] Based on the acquired operating status information of the coal conveyor belt, using an improved YOLOv8 target detection algorithm, obtaining pulley information of the coal conveyor belt; wherein the pulley information may include one or more of the following: bounding box coordinate information, category information, information about the number of pulleys on the left side, and information about the number of pulleys on the right side;

[0087] According to the pulley information of the coal conveyor belt, the difference between the number of pulleys on the left side and the number of pulleys on the right side of the coal conveyor belt is calculated to obtain the difference between the pulleys on both sides of the coal conveyor belt;

[0088] Determining visual offset information of the coal conveyor belt according to a difference between pulleys on both sides of the coal conveyor belt;

[0089] Wherein, the backbone network of the improved YOLOv8 target detection algorithm may be EfficientViT;

[0090] The visual offset information can include the direction and degree of visual offset. In this implementation, the operating status of the coal conveyor belt can be collected by installing a high-definition industrial camera above or to the side of the coal conveyor belt to capture a real-time video stream of the belt's operation. The camera should be positioned to clearly capture the pulleys (such as rollers) on both sides of the belt, and the video frame rate can be set to 3 frames per second or higher to ensure real-time detection.

[0091] In this implementation, the process of determining the visual offset information of the coal conveyor belt based on the pulley difference on both sides of the coal conveyor belt may include:

[0092] If the difference between the pulleys on both sides of the coal conveyor belt is greater than a set first threshold, it is determined that the visual offset direction of the coal conveyor belt is right;

[0093] If the difference between the pulleys on both sides of the coal conveyor belt is less than a set second threshold, it is determined that the visual offset direction of the coal conveyor belt is left;

[0094] If the wheel slip difference on both sides of the coal conveyor belt is within a set first threshold range, it is determined that the coal conveyor belt has no visual deviation;

[0095] If the wheel slip difference on both sides of the coal conveyor belt is within a set second threshold range, it is determined that the visual deviation of the coal conveyor belt is slight;

[0096] If the wheel slip difference on both sides of the coal conveyor belt is within a set third threshold range, it is determined that the visual deviation degree of the coal conveyor belt is moderate;

[0097] If the wheel slip difference on both sides of the coal conveyor belt is within a set fourth threshold range, it is determined that the visual deviation degree of the coal conveyor belt is severe;

[0098] For example, in this implementation, an improved YOLOv8 object detection algorithm model (with EfficientViT as the backbone) is used to process images from the video acquisition module. The model is primarily responsible for detecting pulleys in the image in real time and outputting the bounding box coordinates and category information for each pulley. After processing the image, the model detects all pulleys and sorts them by the x-coordinate of the center point of the target box, dividing them into left and right pulleys, and calculating the number of pulleys on both the left and right sides.

[0099] If the number on the left is greater than the number on the right, the initial judgment is that the offset direction is right;

[0100] If the number on the left is less than the number on the right, the initial judgment is that the offset direction is left;

[0101] If the difference between the two numbers is within a preset threshold (e.g., 10%), it is determined that there is no offset;

[0102] The degree of deviation is preliminarily determined based on the magnitude of the quantitative difference (e.g., 10%-20% is mild, 20%-40% is moderate, and >40% is severe);

[0103] In the above implementation, by introducing the improved YOLOv8 target detection algorithm based on the EfficientViT backbone network into the coal conveyor belt operation status image, high-precision recognition and real-time positioning of the pulley device can be achieved, thereby extracting the number information of the pulleys on the left and right sides of the belt, and based on this, determining the direction and degree of visual offset. As a lightweight and efficient visual Transformer architecture, EfficientViT boasts strong feature representation capabilities and computational efficiency. It can improve the real-time response performance of target detection while maintaining model detection accuracy, making it particularly suitable for rapid inference in edge computing devices deployed in industrial field environments. By identifying the bounding box coordinates and categories of pulleys in images and categorizing them based on their coordinate center points, it can effectively distinguish structural differences on both sides of the belt and quantify the difference in the number of pulleys. This implementation not only improves the sensitivity and accuracy of visual inspection for offset trends but also distinguishes between mild, moderate, and severe offsets by setting grading thresholds (e.g., 10%, 20%, and 40%), providing quantifiable and predictable continuous features in the visual inspection results. Furthermore, by further combining real-time video acquisition with a dynamic frame rate control mechanism, the entire offset identification process is both timely and stable, making it suitable for online monitoring and early warning of belt operation status. This avoids the instability of offset judgment caused by environmental factors such as occlusion and illumination in traditional image recognition methods, thereby improving the accuracy of offset determination for coal conveyor belts and their adaptability to field applications.

[0104] In one implementation, the process of obtaining the physical offset information of the coal conveyor belt based on the sensor measurement values ​​collected by the laser ranging sensors at the symmetrical positions on both sides of the coal conveyor belt in step 2 may include:

[0105] performing difference calculation based on sensor measurement values ​​collected by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt to obtain a laser ranging difference;

[0106] If the laser ranging difference is less than a set distance threshold, it is determined that the coal conveyor belt has not physically deviated;

[0107] If the laser ranging difference is greater than or equal to the distance threshold, it is determined that the coal conveyor belt has physically deviated, and the deflection direction of the coal conveyor belt is used as the physical deflection information of the coal conveyor belt; in this implementation, laser ranging sensors can be installed at symmetrical positions on both sides of the coal conveyor belt (one or more can be installed on each side) to measure the vertical distance between the belt surface or a specific marking point and the sensor in real time, and the sensor installation must ensure that the measurement range covers the width of the normal operation of the belt; and by calculating the difference in the measurement values ​​of the laser ranging sensors on both sides, if the difference is within the preset threshold range, it is considered that the distance between the two sides of the belt is basically balanced; if the difference exceeds the threshold, it indicates that the belt may be deviated, and it can assist in determining the direction of the deflection (the side with the longer distance is the side closer to the sensor);

[0108] For example, in this implementation, the distance data collected by the laser ranging sensor (i.e., sensor measurement value) can also be filtered (e.g., outliers can be removed), and the visually detected pulley information (position, number) and the laser ranging data (distance difference between the two sides) can be fused and analyzed:

[0109] Case 1: Visual detection detects an offset (e.g., a numerical difference exceeding 10%), but the laser ranging data does not exceed the threshold. This triggers a secondary verification mechanism, briefly waiting for the next frame of data to confirm, or combining historical data trend analysis, or temporarily not sending a serious signal and waiting for further confirmation to reduce the possibility of false operations caused by visual false alarms.

[0110] Case 2: If the visual detection of the deviation is positive and the laser ranging data also exceeds the threshold (or both indicate deviation in the same direction), it is determined to be a true deviation. In this case, the deviation degree detected by the visual detection (based on the difference in the number of pulleys) and the more accurate physical distance deviation information provided by the laser ranging are combined to comprehensively determine the final deviation direction and deviation degree (mild, moderate, severe);

[0111] Case 3: No significant deviation is detected visually (the difference is less than 10%), but the laser ranging data continues to exceed the threshold. This may be because the visual feature (pulley) is blocked or difficult to identify. In this case, the laser ranging data should be trusted first, determined to be a deviation, and the event should be recorded for subsequent analysis.

[0112] Therefore, in the above implementation, by combining an improved YOLOv8 target detection algorithm based on the EfficientViT backbone network with symmetrically arranged laser ranging sensors, a multi-source information fusion mechanism for coal conveyor belt deviation detection was constructed, which facilitates improved deviation identification accuracy and stability. For example, in the image processing phase, the improved YOLOv8 model is used to perform high-precision identification of pulley devices in the video stream, obtaining pulley bounding boxes and classification information. The number of pulleys on the left and right sides of the belt is dynamically counted by partitioning based on the target position center point. The deviation direction and degree are then determined by combining the difference in the number of pulleys with a preset threshold, making the visual recognition process highly real-time and capable of classification. At the same time, by symmetrically installing laser ranging sensors on both sides of the coal conveyor belt, the distance data between the belt edge and the sensor is obtained in real time, and the difference between the measured values ​​on the left and right sides is calculated, thereby achieving accurate quantification of the physical offset; this implementation method determines whether there is an offset by setting a distance threshold, and uses the sign of the distance difference to assist in determining the offset direction, effectively making up for the blind spots of visual detection in the presence of occlusion, contamination, or light interference; further, by fusing and analyzing the above two types of detection results, for example, triggering a verification mechanism when there is doubt in a single detection result, enhancing the offset confirmation signal when both determine an offset, and giving priority to a single result when it is credible, it is beneficial to improve the adaptability to complex working conditions.

[0113] In one implementation, the process of performing fusion analysis based on the visual offset information and the physical offset information in step 3 to obtain the offset state information of the coal conveyor belt may include:

[0114] When there is no visual deviation of the coal conveyor belt and the coal conveyor belt is physically deviated, a secondary verification mechanism is triggered to obtain a secondary deviation detection result of the coal conveyor belt, and the secondary deviation detection result is used as the deviation status information of the coal conveyor belt;

[0115] When the visual deviation degree of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt is physically deviated, it is determined that the coal conveyor belt is actually deviated, and the visual deviation degree and physical deviation information of the coal conveyor belt are used as the deviation state information of the coal conveyor belt;

[0116] When the visual deviation of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt has not physically deviated, it is determined that the coal conveyor belt has actually deviated, and the physical deviation information of the coal conveyor belt is used as the deviation state information of the coal conveyor belt;

[0117] In this implementation, when the coal conveyor belt has no visual deviation and the coal conveyor belt has physical deviation, triggering the secondary verification mechanism and obtaining the secondary deviation detection result of the coal conveyor belt may include:

[0118] When the coal conveyor belt has no visual deviation and the coal conveyor belt has physical deviation, retrieving the operation status information of the coal conveyor belt in different time frames;

[0119] According to the operating status information of the coal conveyor belt in different time frames, the improved YOLOv8 target detection algorithm is used to determine the secondary visual offset information of the coal conveyor belt, and the secondary visual offset information is used as the secondary offset detection result; in this implementation method, the mainstream YOLOv8 algorithm is optimized for the coal conveyor belt scenario, and EfficientViT is used as the backbone network, which can improve the operating efficiency of the model while ensuring detection accuracy, and adapt to the real-time requirements of industrial sites.

[0120] Therefore, through the above implementation, a fusion determination mechanism combining visual and physical offset information, combined with a secondary verification strategy based on time frame backtracking, makes the identification of coal conveyor belt offset status more robust and reliable. In this solution, visual offset information is derived from an improved YOLOv8 object detection algorithm, which uses EfficientViT as its backbone network. This algorithm boasts excellent detection accuracy and computational efficiency, enabling real-time extraction of pulley information from images and determination of offset direction and degree. Physical offset information, on the other hand, is based on real-time distance data collected by laser ranging sensors on both sides, providing a quantified edge offset by determining the distance difference. When the two types of information are inconsistent, intelligent judgment is made through the set fusion analysis logic: if there is no visual offset but physical detection shows an offset, a secondary visual inspection process of looking back to the historical time frame is triggered; by calling image frames from the previous period and reusing the target detection model for analysis, it can overcome the misjudgment caused by occlusion, pollution or instantaneous changes in single-frame images, thereby enhancing the continuity and robustness of the detection process in the time dimension. This processing mechanism fully utilizes the complementary advantages of visual and physical data in the time and space dimensions, so that offset judgment does not rely on a single signal source when facing complex working conditions, reducing the possibility of false alarms and missed alarms, and is conducive to improving the recognition ability and response accuracy of real offset events. Moreover, through the combined judgment of slight, moderate, severe and other offset levels and the cross-verification of dual-source data, this implementation method can more accurately output offset status information with directional and severity attributes, providing a reliable basis for subsequent offset control, and helping to ensure the safe and stable operation of the coal transportation system.

[0121] In one implementation, the process of correcting the offset of the coal conveyor belt using a multi-source data fusion algorithm based on the offset state information of the coal conveyor belt in step 4 may include:

[0122] According to the offset state information of the coal conveyor belt, a Kalman filter algorithm is used to predict the offset of the coal conveyor belt to obtain the offset prediction information of the coal conveyor belt;

[0123] Based on the deviation prediction information of the coal conveyor belt and the deviation state information, performing deviation correction on the coal conveyor belt;

[0124] In this implementation, based on the fused data, the belt deviation status can be comprehensively judged, and a control signal including the deviation direction, deviation degree, and credibility level can be generated. The final judgment result (deviation direction, deviation degree, credibility) is packaged into a standard control signal and sent to the mechanical device system (used to receive the signal from the coal conveyor belt deviation detection system and automatically perform the belt deviation adjustment operation according to the signal content). For example, the signal format can include: {"direction":"left / right / none","severity":"minor / moderate / severe / none","confidence":0.0-1.0};

[0125] For example, the above-mentioned mechanical device system is specifically used for:

[0126] Signal reception and analysis: Real-time monitoring of detection control signals, and analysis of information such as the deviation direction, deviation degree, and reliability of the coal conveyor belt;

[0127] Control logic and decision-making: Generate specific mechanical action instructions based on the parsed information and combined with preset control strategies and algorithms;

[0128] Offset direction judgment: If the received direction is "left", the instruction is to adjust the belt to the "right"; if the direction is "right", the instruction is to adjust the belt to the "left"; if the direction is "none", the current state is maintained or fine-tuned;

[0129] Degree of deviation judgment: If the received degree is "mild", the mechanical device is instructed to make a small adjustment with low force; if the degree is "moderate", a medium adjustment with medium force is made; if the degree is "severe", a large adjustment with high force is made. The adjustment force can be preset to multiple levels or dynamically calculated through an algorithm;

[0130] Mechanical execution: Usually installed at the head or middle of the coal conveyor belt, it includes a servo motor-driven mechanical arm, a hydraulic / pneumatic deviation adjustment device, or an electric tensioning / adjustment device. This process accurately performs adjustment actions according to control instructions, changing the position or tension of the belt on the drum, and gradually returning the belt to the correct position.

[0131] State feedback and closed-loop control: After the mechanical device performs an adjustment, the adjustment effect can be fed back through built-in sensors (such as position sensors and force sensors), forming a closed loop with the detection process to achieve more precise automatic adjustment control;

[0132] For example, the application process of the multi-source data fusion algorithm may include:

[0133] State estimation: The belt's offset state (including offset amount, offset speed, etc.) is considered as a dynamic state vector;

[0134] Observation model: The difference in the number of pulleys detected by visual inspection (converted to offset estimation) and the distance difference between the two sides measured by laser ranging are used as observation values;

[0135] Fusion processing: The Kalman filter uses a prediction model (based on the belt's motion inertia and historical state) and the current observation value to recursively estimate the optimal estimate of the belt offset at the current moment. Compared with the direct output of a single sensor, this estimate has lower noise and smoother dynamic response, and can effectively fuse the information of the two sensors, which is conducive to improving the accuracy and robustness of the overall detection. Introducing advanced data fusion algorithms such as Kalman filtering during the detection process, dynamically fusing and estimating the data from vision and laser sensors facilitates the smooth, accurate, and real-time calculation of belt offset, improving the stability and accuracy of the detection results, and providing a high-quality data foundation for subsequent precise automatic deviation adjustment.

[0136] Dynamic correction: The optimal offset estimate output by the filter serves as input to the offset determination and signal generation module, more accurately determining the offset direction and degree, and guiding the mechanical system to make more refined adjustments.

[0137] Therefore, through the above implementation method, a multi-source data fusion mechanism is introduced, and the Kalman filter algorithm is used to jointly model and dynamically estimate the offset state data provided by visual and laser ranging sensors. This can achieve highly stable and high-precision prediction of the coal conveyor belt's offset state, and accordingly generate offset control instructions to drive the mechanical system to perform targeted automatic corrections. Specifically, the offset direction, offset degree, and confidence level of the coal conveyor belt are used as a unified offset state vector. The Kalman filter performs recursive estimation based on historical state evolution and current observations. This can effectively smooth out fluctuation errors in visual recognition and transient anomalies in laser ranging data, improve the ability to continuously track offset trends, and reduce misjudgment interventions caused by single-frame anomalies. At the same time, the offset correction process in this implementation method not only relies on the static judgment results, but also makes dynamic decisions based on the predicted quantities, so that it has feedforward adjustment capabilities, and can implement fine-tuning control before the offset trend reaches a serious level, thereby improving the response speed and flexible control effect of the adjustment; the control signal generated thereby is packaged into a structured format, which can carry complete direction, level and credibility information for accurate analysis by the downstream offset adjustment execution system. The execution system drives the servo motor, hydraulic or pneumatic device and other actuators to adjust according to the corresponding direction and force according to the content of the control signal, and uses the state sensor to feedback the execution effect in real time to form a complete closed-loop control path. This closed-loop offset adjustment scheme, which is achieved through the coordinated implementation of state estimation and dynamic correction, can significantly improve the operating stability and anti-interference ability of the coal conveyor belt system in complex operating environments, ensure that the offset problem is timely and accurately identified and effectively corrected, reduce equipment failure rate, and extend the continuous operation time of the system.

[0138] In summary, the present invention addresses the problems existing in the prior art in the detection of coal conveyor belt deviation, especially the technical defects of false alarms or missed alarms that may occur under specific working conditions such as harsh environments, feature occlusion or sensor failure, resulting in the inability to detect and handle deviations in a timely manner. A method for detecting coal conveyor belt deviation is proposed. Figure 2As shown, this method combines deep learning-based computer vision pulley detection technology with laser ranging technology. By jointly processing and fusing the two types of data, a complete judgment logic mechanism is established: during the detection process, the laser ranging is prioritized to determine whether it exceeds the threshold. If it does not, it is further verified by combining the visual detection results, and a secondary verification mechanism based on historical frames is initiated if necessary. If both the visual and laser data show an offset, it is determined to be a true offset, thereby improving the reliability and accuracy of the offset judgment. Based on the fused offset status information, a multi-source data fusion algorithm is further utilized to generate a control signal including the offset direction, offset degree, and credibility, driving the mechanical device to perform the corresponding offset adjustment action, ultimately achieving a closed-loop control process from offset detection to automatic offset adjustment of the coal conveyor belt. While ensuring detection accuracy, the method of the present invention significantly improves environmental adaptability and anti-interference capabilities, effectively reducing the false alarm rate and missed alarm rate. Overall, the present invention provides a coal conveyor belt offset detection and automatic correction method that integrates computer vision and laser ranging, combined with a multi-source data fusion algorithm. It has high reliability, high precision, and high robustness, and can significantly improve the safety, stability, and operational efficiency of the coal conveyor system.

[0139] Example 2:

[0140] The present invention based on the same inventive concept also provides a coal conveyor belt deviation detection system, the structural composition diagram is as follows Figure 3 As shown, including:

[0141] A visual detection module is used to perform image processing on the acquired operating status information of the coal conveyor belt using a target detection algorithm to obtain visual offset information of the coal conveyor belt;

[0142] a physical detection module, configured to obtain physical offset information of the coal conveyor belt based on sensor measurement values ​​acquired by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt;

[0143] a data fusion module, configured to perform fusion analysis based on the visual offset information and the physical offset information to obtain offset status information of the coal conveyor belt;

[0144] The offset correction module is used to correct the offset of the coal conveyor belt by using a multi-source data fusion algorithm according to the offset status information of the coal conveyor belt.

[0145] In one implementation, the visual detection module may include:

[0146] A pulley detection submodule is configured to obtain pulley information of the coal conveyor belt based on the acquired operating status information of the coal conveyor belt using an improved YOLOv8 object detection algorithm; wherein the pulley information includes one or more of the following: bounding box coordinate information, category information, information about the number of pulleys on the left side, and information about the number of pulleys on the right side;

[0147] A pulley difference calculation submodule is used to calculate the difference between the number of pulleys on the left side and the number of pulleys on the right side of the coal conveyor belt according to the pulley information of the coal conveyor belt, so as to obtain the difference between the pulleys on both sides of the coal conveyor belt;

[0148] A visual offset submodule, configured to determine visual offset information of the coal conveyor belt according to a difference between pulleys on both sides of the coal conveyor belt;

[0149] Among them, the backbone network of the improved YOLOv8 target detection algorithm is EfficientViT;

[0150] The visual offset information includes: a visual offset direction and a visual offset degree.

[0151] In one implementation, the visual offset submodule may include:

[0152] a first direction deviation unit, configured to determine that the visual deviation direction of the coal conveyor belt is right when the difference between the pulleys on both sides of the coal conveyor belt is greater than a set first threshold;

[0153] a second direction deviation unit, configured to determine that the visual deviation direction of the coal conveyor belt is left when the difference between the pulleys on both sides of the coal conveyor belt is less than a set second threshold;

[0154] a no-deviation output unit, configured to determine that the coal conveyor belt has no visual deviation when the wheel slip difference between the two sides of the coal conveyor belt is within a set first threshold range;

[0155] a first degree output unit, configured to determine that the visual deviation degree of the coal conveyor belt is slight when the wheel slip difference between the two sides of the coal conveyor belt is within a set second threshold range;

[0156] a second degree output unit, configured to determine that the visual deviation degree of the coal conveyor belt is moderate when the wheel slip difference on both sides of the coal conveyor belt is within a set third threshold range;

[0157] The third degree output unit is configured to determine that the visual deviation degree of the coal conveyor belt is severe when the wheel slip difference on both sides of the coal conveyor belt is within a set fourth threshold range.

[0158] In one implementation, the physical detection module may include:

[0159] a laser difference calculation submodule, configured to perform difference calculation based on sensor measurement values ​​acquired by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt to obtain a laser ranging difference;

[0160] A physical no-offset submodule, configured to determine that no physical offset has occurred on the coal conveyor belt when the laser ranging difference is less than a set distance threshold;

[0161] The physical offset submodule is used to determine that the coal conveyor belt has physically offset when the laser ranging difference is greater than or equal to the distance threshold, and use the offset direction of the coal conveyor belt as the physical offset information of the coal conveyor belt.

[0162] In one implementation, the data fusion module may include:

[0163] A secondary verification submodule is configured to trigger a secondary verification mechanism when there is no visual deviation of the coal conveyor belt and the coal conveyor belt is physically offset, obtain a secondary offset detection result of the coal conveyor belt, and use the secondary offset detection result as the offset status information of the coal conveyor belt;

[0164] an offset output submodule, configured to determine that the coal conveyor belt has actually offset when the visual offset degree of the coal conveyor belt is slight, moderate, or severe and the coal conveyor belt has physically offset, and to use the visual offset degree and physical offset information of the coal conveyor belt as offset status information of the coal conveyor belt;

[0165] The fuzzy judgment submodule is used to determine that the coal conveyor belt has actually shifted when the visual shift degree of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt has not physically shifted, and to use the physical shift information of the coal conveyor belt as the shift status information of the coal conveyor belt.

[0166] In this implementation, the secondary syndrome module may include:

[0167] a frame information arrangement unit, configured to retrieve the operating status information of the coal conveyor belt in different time frames when the coal conveyor belt has no visual offset and the coal conveyor belt has a physical offset;

[0168] The secondary visual detection unit is used to determine the secondary visual offset information of the coal conveyor belt according to the operating status information of the coal conveyor belt in different time frames using an improved YOLOv8 target detection algorithm, and use the secondary visual offset information as the secondary offset detection result.

[0169] In one implementation, the offset correction module may include:

[0170] A deviation prediction submodule is used to predict the deviation of the coal conveyor belt using a Kalman filter algorithm according to the deviation state information of the coal conveyor belt, so as to obtain the deviation prediction information of the coal conveyor belt;

[0171] The dynamic correction submodule is used to correct the offset of the coal conveyor belt based on the offset prediction information and the offset state information of the coal conveyor belt.

[0172] Example 3:

[0173] like Figure 4 As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0174] The processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to realize the steps of a coal conveyor belt offset detection method in the above embodiment.

[0175] Example 4:

[0176] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in the electronic device for storing programs and data. It can be understood that the storage medium here can include both the built-in storage medium in the electronic device and the extended storage medium supported by the electronic device. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a coal conveyor belt offset detection method in the above embodiment.

[0177] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0178] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0179] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0180] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0181] Finally, 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 its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art may still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims.

Claims

1. A method for detecting deviation of a coal conveyor belt, characterized in that: include: Performing image processing on the acquired operating status information of the coal conveyor belt using a target detection algorithm to obtain visual offset information of the coal conveyor belt; Obtaining physical offset information of the coal conveyor belt based on sensor measurement values ​​collected by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt; Performing a fusion analysis based on the visual offset information and the physical offset information to obtain offset status information of the coal conveyor belt; According to the offset state information of the coal conveyor belt, a multi-source data fusion algorithm is used to correct the offset of the coal conveyor belt.

2. The method according to claim 1, wherein The method of performing image processing on the acquired operating status information of the coal conveyor belt by using a target detection algorithm to obtain visual offset information of the coal conveyor belt includes: Based on the acquired operating status information of the coal conveyor belt, the improved YOLOv8 target detection algorithm is used to obtain pulley information of the coal conveyor belt; wherein the pulley information includes one or more of the following: bounding box coordinate information, category information, information about the number of pulleys on the left side, and information about the number of pulleys on the right side; According to the pulley information of the coal conveyor belt, the difference between the number of pulleys on the left side and the number of pulleys on the right side of the coal conveyor belt is calculated to obtain the difference between the pulleys on both sides of the coal conveyor belt; Determining visual offset information of the coal conveyor belt according to a difference between pulleys on both sides of the coal conveyor belt; Among them, the backbone network of the improved YOLOv8 target detection algorithm is EfficientViT; The visual offset information includes: a visual offset direction and a visual offset degree.

3. The method according to claim 2, wherein Determining the visual offset information of the coal conveyor belt according to the difference between the pulleys on both sides of the coal conveyor belt includes: If the difference between the pulleys on both sides of the coal conveyor belt is greater than a set first threshold, it is determined that the visual offset direction of the coal conveyor belt is right; If the difference between the pulleys on both sides of the coal conveyor belt is less than a set second threshold, it is determined that the visual offset direction of the coal conveyor belt is left; If the wheel slip difference on both sides of the coal conveyor belt is within a set first threshold range, it is determined that the coal conveyor belt has no visual deviation; If the wheel slip difference on both sides of the coal conveyor belt is within a set second threshold range, it is determined that the visual deviation of the coal conveyor belt is slight; If the wheel slip difference on both sides of the coal conveyor belt is within a set third threshold range, it is determined that the visual deviation degree of the coal conveyor belt is moderate; If the wheel slip difference on both sides of the coal conveyor belt is within a set fourth threshold range, it is determined that the visual deviation degree of the coal conveyor belt is serious.

4. The method according to claim 3, wherein The physical offset information of the coal conveyor belt is obtained based on the sensor measurement values ​​collected by the laser ranging sensors at the symmetrical positions on both sides of the coal conveyor belt, including: performing difference calculation based on sensor measurement values ​​collected by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt to obtain a laser ranging difference; If the laser ranging difference is less than a set distance threshold, it is determined that the coal conveyor belt has not physically deviated; If the laser ranging difference is greater than or equal to the distance threshold, it is determined that the coal conveyor belt has physically deviated, and the deflection direction of the coal conveyor belt is used as the physical deflection information of the coal conveyor belt.

5. The method according to claim 4, wherein The performing of a fusion analysis based on the visual offset information and the physical offset information to obtain the offset state information of the coal conveyor belt includes: When there is no visual deviation of the coal conveyor belt and the coal conveyor belt is physically deviated, a secondary verification mechanism is triggered to obtain a secondary deviation detection result of the coal conveyor belt, and the secondary deviation detection result is used as the deviation status information of the coal conveyor belt; When the visual deviation degree of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt is physically deviated, it is determined that the coal conveyor belt is actually deviated, and the visual deviation degree and physical deviation information of the coal conveyor belt are used as the deviation state information of the coal conveyor belt; When the visual deviation of the coal conveyor belt is slight, moderate or severe, and the coal conveyor belt has not physically deviated, it is determined that the coal conveyor belt has actually deviated, and the physical deviation information of the coal conveyor belt is used as the deviation status information of the coal conveyor belt.

6. The method according to claim 5, wherein When the coal conveyor belt has no visual deviation and the coal conveyor belt has physical deviation, a secondary verification mechanism is triggered to obtain a secondary deviation detection result of the coal conveyor belt, including: When the coal conveyor belt has no visual deviation and the coal conveyor belt has physical deviation, retrieving the operation status information of the coal conveyor belt in different time frames; According to the operating status information of the coal conveyor belt in different time frames, the improved YOLOv8 target detection algorithm is used to determine the secondary visual offset information of the coal conveyor belt, and the secondary visual offset information is used as the secondary offset detection result.

7. The method according to claim 1, wherein The method of correcting the offset of the coal conveyor belt by using a multi-source data fusion algorithm according to the offset state information of the coal conveyor belt comprises: According to the offset state information of the coal conveyor belt, a Kalman filter algorithm is used to predict the offset of the coal conveyor belt to obtain the offset prediction information of the coal conveyor belt; Based on the deviation prediction information and the deviation status information of the coal conveyor belt, the deviation of the coal conveyor belt is corrected.

8. A coal conveyor belt deviation detection system, characterized in that: include: A visual detection module is used to perform image processing on the acquired operating status information of the coal conveyor belt using a target detection algorithm to obtain visual offset information of the coal conveyor belt; a physical detection module, configured to obtain physical offset information of the coal conveyor belt based on sensor measurement values ​​acquired by laser ranging sensors at symmetrical positions on both sides of the coal conveyor belt; a data fusion module, configured to perform fusion analysis based on the visual offset information and the physical offset information to obtain offset status information of the coal conveyor belt; The offset correction module is used to correct the offset of the coal conveyor belt by using a multi-source data fusion algorithm according to the offset status information of the coal conveyor belt.

9. The system according to claim 8, wherein The visual inspection module comprises: A pulley detection submodule is configured to obtain pulley information of the coal conveyor belt based on the acquired operating status information of the coal conveyor belt using an improved YOLOv8 object detection algorithm; wherein the pulley information includes one or more of the following: bounding box coordinate information, category information, information about the number of pulleys on the left side, and information about the number of pulleys on the right side; A pulley difference calculation submodule is used to calculate the difference between the number of pulleys on the left side and the number of pulleys on the right side of the coal conveyor belt according to the pulley information of the coal conveyor belt, so as to obtain the difference between the pulleys on both sides of the coal conveyor belt; A visual offset submodule, configured to determine visual offset information of the coal conveyor belt according to a difference between pulleys on both sides of the coal conveyor belt; Among them, the backbone network of the improved YOLOv8 target detection algorithm is EfficientViT; The visual offset information includes: a visual offset direction and a visual offset degree.

10. The system according to claim 9, wherein: The visual offset submodule includes: a first direction deviation unit, configured to determine that the visual deviation direction of the coal conveyor belt is right when the difference between the pulleys on both sides of the coal conveyor belt is greater than a set first threshold; a second direction deviation unit, configured to determine that the visual deviation direction of the coal conveyor belt is left when the difference between the pulleys on both sides of the coal conveyor belt is less than a set second threshold; a no-deviation output unit, configured to determine that the coal conveyor belt has no visual deviation when the wheel slip difference between the two sides of the coal conveyor belt is within a set first threshold range; a first degree output unit, configured to determine that the visual deviation degree of the coal conveyor belt is slight when the wheel slip difference between the two sides of the coal conveyor belt is within a set second threshold range; a second degree output unit, configured to determine that the visual deviation degree of the coal conveyor belt is moderate when the wheel slip difference on both sides of the coal conveyor belt is within a set third threshold range; The third degree output unit is configured to determine that the visual deviation degree of the coal conveyor belt is severe when the wheel slip difference on both sides of the coal conveyor belt is within a set fourth threshold range.

Images