Camera-chain-based inspection system and method for long-distance belt conveyors
By using a camera chain system to stitch images and rotate and adjust the long-distance belt conveyor, efficient and comprehensive inspection of the entire conveyor belt is achieved, solving the problems of low inspection efficiency and insufficient comprehensive inspection capabilities, and providing functions for damage tracking and coal quantity distribution detection.
Patent Information
- Application Number
- PCT/CN2025/105302
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-14
- Filing Date
- 2025-06-30
- Publication Date
- 2026-11-19
Smart Images

Figure CN2025105302_19112026_PF_FP_ABST
Abstract
Description
A long-distance belt conveyor inspection system and method based on camera chain. Technical Field
[0001] This invention relates to a long-distance belt conveyor inspection system and method based on a camera chain, belonging to the field of long-distance belt conveyor inspection technology. Background Technology
[0002] As a crucial transportation equipment for materials such as coal, the normal operation of belt conveyors is of great significance for improving enterprise production efficiency and promoting safe production in coal mines. To achieve safe and efficient operation of belt conveyors, inspections are typically conducted using methods such as manual labor, inspection robots, and cameras to promptly detect and address faults. Manual inspection suffers from drawbacks such as low efficiency, over-reliance on the experience of inspectors, and a high rate of missed inspections. Inspection robots can perform detailed and comprehensive checks on belt conveyors. Current technologies analyze images collected by inspection robots along the conveyor belt path to check the conveyor's operating status. However, due to current technological limitations, inspection robots suffer from short single-inspection distances, long charging times, and high construction costs, severely impacting inspection efficiency. Cameras can provide long-term, stable monitoring of belt conveyors. Current technologies acquire conveyor belt images through image acquisition devices and use algorithms to analyze and detect damage such as tearing, wear, and cracking. However, this method only provides single-function monitoring of specific areas and lacks the ability to comprehensively inspect the entire belt conveyor. Summary of the Invention
[0003] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide a long-distance belt conveyor inspection system and method based on camera chain, which can perform efficient, comprehensive and integrated inspection of long-distance belt conveyors.
[0004] Prior to this invention, a long-distance belt conveyor inspection system based on a camera chain is provided, comprising a plurality of cameras arranged around the long-distance belt conveyor, a plurality of pan-tilt units for rotating the cameras, a plurality of pan-tilt motors for rotating the pan-tilt units, a plurality of markers for stitching images captured by the cameras, and a computer. The cameras are fixed on the upper surface of the pan-tilt units and rotate with the pan-tilt units. The computer is communicatively connected to the cameras, and the markers are distributed among the cameras.
[0005] Preferredly, the method of deploying cameras includes: deploying cameras in key areas including the radius turning points of the long-distance belt conveyor, the intersection of the long-distance belt conveyor and the pedestrian road, and the uphill and downhill sections of the long-distance belt conveyor; and deploying cameras along the non-key areas between two key areas at a preset maximum distance.
[0006] Prior to this, a long-distance belt conveyor inspection method based on a camera chain utilizes the aforementioned long-distance belt conveyor inspection system based on a camera chain to perform the following steps: S1. Detecting damage in the image captured by the camera using a computer; Comparing the damage location with the center position of the corresponding image using the computer; If the damage location is located to the left of the image center, the pan-tilt unit controls the camera to rotate to the left by a preset degree; If the damage location is located to the right of the image center, the pan-tilt unit controls the camera to rotate to the right by a preset degree; S2. Determining whether the damage location and the center position of the corresponding image are on the same vertical plane; If not, determining whether the current angle of the camera exceeds the preset optimal observation angle; If the current angle of the camera does not exceed the preset optimal observation angle, repeating step S1; otherwise, proceeding to the next step; S3. If the current angle of the camera exceeds the preset optimal observation angle, using the pan-tilt unit to move the camera back to the preset normal working position, and using the adjacent camera to perform step S.
[0007] Prioritizes the following: The first optimal viewing angle for a camera located on a straight section of the conveyor belt in a long-distance belt conveyor is: Where d is the distance between the camera and the longitudinal centerline of the conveyor belt of the long-distance belt conveyor; S is the maximum length of the conveyor belt at which the image clarity meets the preset detection requirements when the distance between the camera and the longitudinal centerline of the conveyor belt is d; the second optimal observation angle of the camera located at the turning section of the conveyor belt is: Where R is the turning radius of the turning section of the conveyor belt of the long-distance belt conveyor.
[0008] Prioritizes using a computer to stitch together images captured by cameras, including: extracting features of landmarks and human landmarks from images captured by all cameras using a computer; matching features of landmarks in images captured by adjacent cameras to identify the same features in adjacent images; estimating the transformation matrix between images based on the same features in the matched adjacent images; and aligning and fusing the images using the transformation matrix based on human landmark features.
[0009] Prior to this, reflective materials are used for the markers, and images with the markers are captured using a camera. The markers in the images are then used as image registration features by a computer. Based on the image registration features, multiple images are stitched together by a computer to obtain a complete image of the long-distance belt conveyor.
[0010] Firstly, a unique ID number is preset for each camera using a computer; when staff conduct patrols, the computer retrieves images of the corresponding locations using the camera's ID number.
[0011] First, use a computer to determine whether the images captured by the camera are overloaded or even overloaded; if so, use a computer to issue a safety warning.
[0012] Preferably, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method described in any of the first aspects.
[0013] Preferably, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects.
[0014] The beneficial effects achieved by this invention are as follows: Compared with traditional inspection robots for monitoring long-distance belt conveyors, this invention uses a camera chain as a conveyor belt information acquisition device, enabling long-term stable monitoring of the conveyor belt and solving the problem of low inspection efficiency when using inspection robots. Compared with the traditional method of monitoring conveyor belts using cameras, this invention uses a camera chain to comprehensively inspect the entire long-distance belt conveyor. Image stitching technology can obtain images of the entire conveyor belt, and the rotation of the cameras can track and confirm any damage found. Furthermore, this invention can perform comprehensive inspection of the entire conveyor belt, including damage detection and coal distribution detection. Attached Figure Description
[0015] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 is a structural diagram of some embodiments of this application; Figure 2 is a flowchart of some embodiments of this application; Figure 3 is a schematic diagram of the optimal observation angle of the camera at the straight section of the conveyor belt in some embodiments of this application; Figure 4 is a schematic diagram of the optimal observation angle of the camera at the turning section of the conveyor belt in some embodiments of this invention.
[0017] The following are the meanings of the labels in the attached diagram: 1-Computer; 2-Long-distance belt conveyor; 3-Marker; 4-Camera; 5-Gimbal; 6-Gimbal motor; 7-Conveyor belt. Detailed Implementation
[0018] Referring to Figure 1, this application discloses a long-distance belt conveyor inspection system based on a camera chain, comprising several cameras 4 arranged around the long-distance belt conveyor 2, several pan-tilt units 5 for rotating the cameras 4, several pan-tilt motors 6 for rotating the pan-tilt units 5, several markers 3, and a computer 1. The markers 3 are distributed between two adjacent cameras 4 and near the conveyor belt 7. The cameras 4 are fixed on the upper surface of the pan-tilt units 5 and rotate with the pan-tilt units 5. The computer 1 is communicatively connected to the cameras 4. The markers 3 are distributed in fixed positions in the images captured by the cameras 4.
[0019] The marker 3 is in the shape of a checkerboard or a pentagon. In the image captured by the camera 4, the marker 3 is distributed in four positions: upper left, lower left, upper right, and lower right. The surface of the marker 3 is made of reflective material to ensure that it can be clearly displayed in the captured image. The gimbal 5 drives the rotor to rotate relative to the stator through the gimbal motor 6. The computer 1 is used to store the damage detection model, the coal quantity detection model, and the image stitching algorithm, process and display the captured images, and at the same time, the gimbal 5 controls the action of the gimbal motor 6 to realize the tracking of the damage position of the conveyor belt 7 by the camera 4.
[0020] In this embodiment, the method of arranging cameras 4 along the long-distance belt conveyor 2 includes: arranging cameras 4 in key areas including the radius turning points of the long-distance belt conveyor 2, the intersection areas of the long-distance belt conveyor 2 and pedestrian roads, and the uphill and downhill sections of the long-distance belt conveyor 2; for other non-key areas, to save costs, they are arranged according to the maximum distance principle, that is, the area that one camera 4 can cover is x meters, and the distance along the line of other non-key areas between two key areas is L meters. According to the maximum distance principle, cameras 4 should be arranged at equal intervals of L / x+1 in the non-key areas, where L / x represents the integer part of L / x. Using L / x+1 cameras 4 ensures that there are no blind spots in the monitoring.
[0021] In this embodiment, the marker 3 near the conveyor belt 7 in the long-distance belt conveyor 2 is made of reflective material, which can be clearly displayed in the image captured by the camera 4. Due to the limitations of the dim environment in underground coal mines, image stitching algorithms may not be able to extract enough features for image registration, resulting in inaccurate image stitching. The special shape of the marker provides image registration features for the image stitching algorithm, helping to achieve accurate image stitching.
[0022] In this embodiment of the application, each camera 4 has its own ID number. When staff need to patrol a specific area, they can retrieve the image information of the corresponding location through the ID number of the camera 4.
[0023] In this embodiment, when computer 1 detects damage to conveyor belt 7, in order to better observe the damage, camera 4 needs to rotate to display the damage location in the center of the image and track the damage location. The specific steps are as follows: S1. Computer 1 detects damage in the image captured by camera 4; Computer 1 compares the damage location with the center position of the corresponding image. If the damage location is on the left side of the image center, the gimbal controls camera 4 to rotate to the left by a preset degree; if the damage location is on the right side of the image center, the gimbal controls camera 4 to rotate to the right by a preset degree; S2. Determine whether the damage location is on the same vertical plane as the center position of the corresponding image. If not, determine whether the current angle of camera 4 exceeds the preset optimal observation angle; If the current angle of camera 4 does not exceed the preset optimal observation angle, repeat step S1; otherwise, proceed to the next step; S3. If the current angle of camera 4 exceeds the preset optimal observation angle, the gimbal 5 drives camera 4 back to the preset normal working position, and the adjacent camera 4 executes step S1.
[0024] In this embodiment of the application, as shown in Figure 3, the specific value of the optimal observation angle is determined in two cases: when the camera 4 is located on the straight section of the conveyor belt 7, the first optimal observation angle is... Where d is the distance between camera 4 and the longitudinal centerline of conveyor belt 7; S is the maximum length of conveyor belt 7 when the image clarity meets the detection requirements at a distance d from camera 4 to the longitudinal centerline of conveyor belt 7; when camera 4 is located at a turning section of conveyor belt 7, the optimal observation angle β is obtained by using the law of cosines within triangle △ABC: Combining the above formula, we can obtain the second optimal observation angle for the 7-turn section of the conveyor belt: Where R is the turning radius of the turning section of conveyor belt 7; as shown in Figure 4, segment DE is the part of the image that meets the detection requirements when the turning section of conveyor belt 7 is transformed into a straight line, and segment CD is an arc with point B as the center and r as the radius. Since the distances of points C and D from camera 4 are equal, point C can be regarded as the boundary point of the turning section of conveyor belt 7 that meets the detection requirements in the image.
[0025] In this embodiment, the image stitching method uses a feature registration-based stitching algorithm. First, features are extracted from adjacent images, including manually set landmark features and other features in the images. Second, feature matching is performed between adjacent images to identify and match the same features in adjacent images. Then, the transformation matrix between images is estimated through the matched features. Finally, the images are aligned and fused using the transformation matrix.
[0026] Because it's difficult to guarantee perfect synchronization when images are captured by adjacent cameras 4 (meaning the stitched images are not taken at the same time but have a certain time interval), features such as the coal flow profile change between the two images. This leads to errors when estimating the transformation matrix using matched features. Therefore, this invention incorporates a human-made marker feature. This feature is stationary relative to the conveyor belt 7, and even if there's a time interval between images captured by adjacent cameras 4, this feature remains unchanged in the image, thus correcting and reducing errors in estimating the transformation matrix.
[0027] In this embodiment, as shown in Figure 2, the coal quantity detection model categorizes coal quantity into five states: empty, lightly loaded, medium-loaded, heavily loaded, and overloaded. It identifies the coal quantity in images captured by camera 4, determining if there is an extremely uneven coal quantity distribution. If so, a safety warning is issued, and the coal quantity distribution on the long-distance conveyor belt 7 is integrated. Specifically, during coal quantity analysis, it is not necessary to process every image captured by camera 4; analysis can be performed every few images. Furthermore, due to the continuous transport of the belt conveyor, high-frequency detection of the conveyor belt 7 is unnecessary; images can be updated every few seconds. When localized areas on the conveyor belt 7 are heavily loaded or even overloaded while other areas are lightly loaded or empty, the conveyor belt 7 may experience a severe imbalance in tension distribution. Computer 1 identifies this situation and issues a safety warning to the personnel.
[0028] In this embodiment of the application, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the methods described above.
[0029] In this application embodiment, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.
[0030] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
[0031] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention described herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not invented herein. The specification and examples are to be considered exemplary only.
[0032] The above specific embodiments further illustrate the purpose, technical solution and beneficial effects of this application. It should be understood that the above are only specific embodiments of this application and are not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of this application should be included within the scope of protection of this application.
Claims
1. A long distance belt conveyor inspection system based on camera chain, characterized in that, It includes several cameras (4) set around a long-distance belt conveyor (2), several pan-tilt units (5) used to drive the cameras (4) to rotate, several pan-tilt motors (6) used to drive the pan-tilt units (5) to rotate, several markers (3) used to realize the stitching of images acquired by the cameras (4) and a computer (1). The cameras (4) are fixed on the upper surface of the pan-tilt units (5) and the cameras (4) follow the rotation of the pan-tilt units (5). The computer (1) is connected to the cameras (4) for communication. Several markers (3) are distributed among several cameras (4).
2. A long distance belt conveyor inspection system based on camera chain according to claim 1, characterized in that, Methods for arranging cameras (4) include: Cameras (4) are installed in key areas including the radius turning position of the long-distance belt conveyor (2), the intersection of the long-distance belt conveyor (2) and the pedestrian road, and the uphill and downhill positions of the long-distance belt conveyor (2); Cameras are placed along the non-key area between the two key areas at a preset maximum distance (4).
3. A long distance belt conveyor inspection method based on camera chain, characterized in that, Using the long-distance belt conveyor inspection system based on a camera chain as described in claim 1 or 2, the following is performed: S1. Damage is detected in the image captured by the camera (4) using a computer (1); The computer (1) compares the damaged location with the center of the corresponding image. If the damaged location is on the left side of the image center, the gimbal (5) controls the camera (4) to rotate to the left by a preset degree; if the damaged location is on the right side of the image center, the gimbal (5) controls the camera (4) to rotate to the right by a preset degree. S2. Determine whether the location of the damage is on the same vertical plane as the center of the corresponding image. Otherwise, determine whether the current angle of the camera (4) exceeds the preset optimal observation angle. If the current angle of the camera (4) does not exceed the preset optimal observation angle, repeat step S1; otherwise, proceed to the next step. S3. If the current angle of the camera (4) exceeds the preset optimal observation angle, use the gimbal (5) to drive the camera (4) back to the preset normal working position, and use the adjacent camera (4) of the camera (4) to execute step S1.
4. The long distance belt conveyor inspection method based on camera chain according to claim 3, characterized in that, The first optimal viewing angle of the camera (4) located at a straight section of the conveyor belt (7) in a long distance belt conveyor (2) is: Where d is the distance between the camera (4) and the longitudinal center line of the conveyor belt (7); S is the maximum length of the conveyor belt when the image clarity meets the preset detection requirements when the distance between the camera (4) and the longitudinal center line of the conveyor belt (7) is d. The second optimal viewing angle of the camera (4) located at the turning section of the conveyor belt (7) is: Where R is the turning radius of the turning section of the conveyor belt (7).
5. The long distance belt conveyor inspection method based on camera chain according to claim 3, characterized in that, Using a computer (1) to stitch together images captured by a camera (4), including: The computer (1) was used to extract the features of the markers (3) and human markers in the images captured by all cameras (4); Match the features of the markers (3) in the images captured by the adjacent cameras (4) to identify the same features in the adjacent images; Based on the same feature in the matched adjacent images, the transformation matrix between the images is estimated; Based on human-made marker features, images are aligned and fused using a transformation matrix.
6. The long distance belt conveyor inspection method based on camera chain according to claim 3, characterized in that, The marker (3) is made of reflective material, and the image of the marker (3) is captured by the camera (4); Using a computer (1), the markers (3) in the image are used as image registration features; Based on image registration features, the computer (1) is used to splice multiple images to obtain a complete long-distance belt conveyor image.
7. The long distance belt conveyor inspection method based on camera chain according to claim 3, characterized in that, The computer (1) is used to preset a unique ID number for each camera (4). When the staff conducts a tour, the computer (1) is used to call up the image of the corresponding position through the ID number of the camera (4).
8. The long distance belt conveyor inspection method based on camera chain according to claim 3, characterized in that, The computer (1) is used to determine whether the image captured by the camera (4) is overloaded or even overloaded, and if so, the computer (1) is used to issue a safety warning.
9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the steps of the method of any one of claims 3-8 when executing the program.
10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program implements the steps of the method of any one of claims 3-8 when executed by the processor.
