Mining scraper conveyor on-line round link pitch measurement method

Abstract

This invention discloses an online method for measuring the pitch of a circular link chain in a mining scraper conveyor, belonging to the field of chain breakage monitoring in mining scraper conveyors. The measurement method includes: installation and debugging of a camera device to obtain basic image parameters; image processing and pitch calculation: acquiring the original image, optical distortion correction, perspective distortion correction, target detection and instance segmentation, calculating the pixel coordinates of the centroid of the horizontal ring and the abscissas of the left and right boundaries of the scraper, calculating the spatial position of the scraper and the circular link chain in a two-dimensional model, converting the pixel coordinates of the centroid of the horizontal ring to physical coordinates of the calibration plane, calculating the pitch of the circular link chain under tension, considering the horizontal ring pitch correction coefficient when the chain is suspended, and finally calculating the compensated and corrected circular link chain pitch. To achieve high-precision online measurement of the circular link chain pitch, this invention provides a method for acquiring and analyzing two-dimensional images using a monocular camera; the pitch measurement is less affected by the movement of the circular link chain, achieving an online measurement accuracy of ±1mm.

Description

Technical Field

[0001] This invention belongs to the field of chain breakage monitoring for mining scraper conveyors, and relates to an online method for measuring the pitch of circular link chains in mining scraper conveyors. Background Technology

[0002] Scraper conveyors are one of the key pieces of equipment in modern underground coal mining operations, and their safe operation is a crucial factor determining whether a coal mine can operate normally. With the development of modern underground coal mines, scraper conveyors are being laid for increasingly longer lengths and have higher conveying power. However, the circular link chain is the least reliable key component in a scraper conveyor, making online monitoring of the circular link chain essential for ensuring the safe operation of the scraper conveyor.

[0003] The primary failure mode of circular link chains is chain breakage. Chain breakage is caused by two main factors: firstly, wear and plastic deformation at the meshing point between the chain and sprocket teeth; and secondly, the formation and development of micro-cracks during chain use. Both of these factors lead to an increase in the chain pitch during service. When the pitch increases to a certain extent, chain breakage occurs, resulting in significant economic losses and risks to equipment and personnel safety. Therefore, it is necessary to implement early warning systems for chain breakage accidents in scraper conveyors to reduce unauthorized downtime.

[0004] To enable early warning of chain breakage accidents, since pitch is one of the key indicators characterizing damage to circular links, online measurement of pitch can effectively predict the occurrence of chain breakage accidents.

[0005] For the current equipment status in coal mine production, the pitch specifications of the circular link chains used in a large number of equipment range from 126mm to 170mm. According to the requirements of the production suppliers, it is generally recommended to replace them when the elongation is 3%, that is, when the elongation is 4mm to 5mm, they need to be replaced. Considering that the tolerance during the production of circular link chains is ±1.5mm, and taking into account the preparation cycle for replacement, when the measurement accuracy is controlled below ±1mm, the measured data can meet the accuracy requirements for the maintenance and upkeep of circular link chains.

[0006] In the prior art, Chinese patent CN223632431U discloses a structure for mounting a camera on a camera headstock; Chinese patent CN118387559A proposes a method for identifying chain breakage through image capture and performing shutdown protection; Chinese patent CN115690517A proposes using an image moving target detection algorithm to detect chain breakage; Chinese patent CN119117584A proposes a method for identifying chain breakage through image capture using AI; and Chinese patent CN115818158A proposes a method for determining chain breakage by capturing images of the engagement state of the circular link chain and sprocket. These patents all contribute to the method of using a monocular camera for online machine vision to acquire image information of the circular link chain and using this image information for circular link chain status analysis. However, due to the following reasons... The following limitations prevent the circular link chain measurement from meeting accuracy requirements: 1. Circular link chains easily adhere to residual coal, making it difficult to achieve the required accuracy for the distance between two measurement points in the image directly acquired by a monocular camera; 2. During movement, tension changes cause the circular link chain to dangle from the bottom of the machine slot, resulting in a suspended chain. The height of this suspended chain changes with tension, leading to dimensional discrepancies and perspective distortion in the image captured by the camera, affecting measurement accuracy; 3. Due to the limiting mechanism of the scraper conveyor's running trough, the scraper's trajectory follows the bottom of the tail trough, but the circular link chain remains suspended in the air. As the scraper moves, the suspended chain also moves, causing the object distance and the angle between the captured image and the imaging plane to constantly change, severely impacting measurement accuracy. Clearly, the visual monitoring in the aforementioned patent literature can only monitor the state after chain breakage and more obvious faults such as scraper skewing and chain stacking. It cannot meet the accuracy requirements for pitch measurement.

[0007] Chinese patent CN119429571A proposes using a binocular camera to acquire 3D image information of the circular link chain as a model for monitoring the circular link chain of a scraper conveyor. This patent is not only hardware-complex, but also suffers from long 3D modeling calculation times, making it difficult to adapt to the computational speed required for online operation and online modeling. Therefore, this technology is difficult to implement in production practice. Thus, there is an urgent need for a method to monitor the pitch of the circular link chain in a mining scraper conveyor online. Summary of the Invention

[0008] To achieve high-precision online measurement of circular link chain pitch, this invention provides an online method for measuring the pitch of circular link chains in mining scraper conveyors. The method employs a monocular camera to acquire two-dimensional images and perform analysis and calculations. The pitch measurement is minimally affected by the movement of the circular link chain, achieving an online measurement accuracy of ±1mm.

[0009] This invention provides a method for online measurement of the pitch of a circular link chain in a mining scraper conveyor, comprising the following steps:

[0010] Step 1: Equipment Installation and Debugging

[0011] The camera is installed above the middle circular chain at the tail of the scraper conveyor, with the camera's optical axis perpendicular to the bottom arc of the tail groove. The camera's field of view covers 8 to 10 circular chains at the sprocket outlet.

[0012] The camera was adjusted to obtain the average millimeter-to-pixel ratio s, the camera intrinsic parameter matrix K, and the distortion coefficients. Calibration plane homography matrix H, image acquisition time interval The scale conversion factor a, and the position parameters of the crosshair points B and C and the camera optical center P in the two-dimensional model;

[0013] Step 2: Image Processing and Pitch Calculation

[0014] (1) Obtain the original image

[0015] Start the camera to continuously acquire raw images;

[0016] Make four cross-shaped markers on the upper surface of the tail section of the camera within the camera's field of view. These are designated as the first marker, the second marker, the third marker, and the fourth marker. The arrangement of the four cross-shaped markers should meet the following requirements: the four cross-shaped markers should form a rectangle in physical space; the center of the rectangle should coincide with the position corresponding to the main point of the camera as much as possible; and the length of the rectangle should be consistent with the direction of the circular chain's movement.

[0017] (2) Optical distortion correction

[0018] Using the camera intrinsic parameter matrix K and the distortion coefficient d, optical distortion correction is performed on the acquired original image to obtain an optically distorted image;

[0019] (3) Correction of perspective distortion

[0020] Using the calibrated plane homography matrix H, perspective distortion correction is performed on the optically distorted image to obtain a perspective-distorted image;

[0021] (4) Target detection and instance segmentation

[0022] The perspective-de-distorted image is input into the instance segmentation model. The complete flat ring in the image is divided into upper and lower halves for detection and pixel-level segmentation, generating blue mask regions for the upper half and lower half of the flat ring respectively. At the same time, the scraper in the image is also detected and segmented at the pixel level, generating blue mask regions for the scraper, and the corresponding binary masks are output to obtain the target detection image.

[0023] (5) Calculate the pixel coordinates of the centroid of the flat ring and the x-coordinates of the pixels at the left and right boundaries of the scraper.

[0024] In the target detection image, the binary masks corresponding to the upper half of the flat ring, the lower half of the flat ring, and the scraper are obtained respectively. The target detection image contains two circular chain chains, and the flat rings on each circular chain are independently numbered. Let the binary mask of the upper half of the k-th flat ring in the image be... The lower half of the binary mask is The binary mask for the i-th scraper from the left in the image is: ,in When the binary mask value is 1, it represents a pixel. It belongs to the target area; otherwise, it does not belong to the area.

[0025] (6) Calculate the spatial positions of the scraper and the circular chain in the two-dimensional model based on the pixel coordinates of the left and right boundaries of the scraper identified in the image:

[0026] Using the horizontal coordinates of the left and right boundaries of the first scraper in the image obtained by step (5) above , Convert them to the horizontal coordinates in the calibration plane coordinate system, respectively. , ;

[0027] (7) Convert the pixel coordinates of the centroid of the flat ring into the physical coordinates of the calibration plane.

[0028] The pixel coordinates of the k-th centroid of the flat ring in the image obtained in step (5) above. Convert to physical coordinates on the calibration plane using the following formula. .

[0029]

[0030] In the formula: , , , , , , .

[0031] Where, when the pixel x-coordinate u of the centroid of the k-th flat ring in the image k smaller than the left boundary pixel value of the scraper hour, , , When the pixel x-coordinate u of the centroid of the k-th flat ring in the image k The pixel value greater than the right boundary of the scraper hour, , , d is the diameter of the circular chain.

[0032] (8) Calculate the pitch of the circular link chain under tension.

[0033] Using the physical coordinates of the kth and (k+1)th centroids of the horizontal rings obtained in step (7) above on the calibration plane , Calculate the pitch P of the circular link chain under tension using the following formula. k :

[0034] ;

[0035] When there is no scraper between two adjacent flat rings, the distance between the centers of gravity of the two flat rings is twice the pitch of the circular chain. When there is a scraper between two adjacent flat rings, the distance between the centers of gravity of the two flat rings is four times the pitch of the circular chain. ;

[0036] (9) Calculate the horizontal link distance correction factor when considering chain suspension.

[0037] The theoretical pixel distance between the centers of gravity of the upper and lower halves of the k-th flat ring in the image under tautness is calculated using the following formula. Combined with the pixel distance between the centroids of the upper and lower halves of the k-th flat ring actually measured in the image, The k-th parallel distance correction coefficient r is obtained. k And use this coefficient to determine the pitch P of the circular chain obtained in step (8) under tension. k Compensation corrections are made to improve the accuracy of the circular link chain pitch measurement results.

[0038]

[0039]

[0040] In the formula: , , , , , , , , , , , ; ;

[0041] Among them, P 1k P 2k q represents the physical coordinates of the centroids of the upper and lower halves of the k-th flat ring in the image on the calibration plane; 1k q 2kThese are the pixel coordinates of the centroids of the upper and lower halves of the k-th flat ring, respectively; when the x-coordinate of the centroid of the k-th flat ring in the image is u... k smaller than the left boundary pixel value of the scraper hour, , , When the pixel x-coordinate u of the centroid of the k-th flat ring in the image k The pixel value greater than the right boundary of the scraper hour, , , ; d is the diameter of the circular chain; s y This is the standard physical distance between the centers of gravity of the upper and lower halves of the flat ring.

[0042] (10) Calculate the corrected pitch of the circular chain.

[0043] The pitch P of the circular chain obtained in step (8) under tension is k The distance between the centroids of the kth and (k+1)th horizontal rings is obtained, therefore it is subject to the correction coefficient r of the distance between the kth and (k+1)th horizontal rings. k r k+1 The combined effects of these factors. Therefore, the tension coefficients of the k-th and (k+1)-th flat rings are averaged, and the pitch P of the circular chain under tension is calculated according to the formula. k Perform compensation and correction, and calculate the final circular chain pitch P. k final :

[0044] .

[0045] The beneficial effects of this invention are:

[0046] (1) This invention only requires one camera, power supply, and one computer; it is easy to use;

[0047] (2) The pitch measurement is less affected by the movement of the circular chain and can achieve an online measurement accuracy of ±1mm;

[0048] (3) The pitch measurement method is not sensitive to residual coal and slurry, and has high measurement accuracy. Attached Figure Description

[0049] Figure 1 Diagram showing the connection between the camera and the scraper conveyor;

[0050] Figure 2 This is a structural diagram of the camera and its support.

[0051] Figure 3 This is a schematic diagram of a two-dimensional model of a circular link chain under tension.

[0052] Figure 4 A schematic diagram of a two-dimensional model of a taut circular chain when the right boundary x-coordinate of the scraper is equal to the right boundary x-coordinate of the image.

[0053] Figure 5 This is a schematic diagram of a two-dimensional model of the taut circular chain when the second imaging ray intersects with the lower surface of the chain pressure block.

[0054] Figure 6 This is a schematic diagram of a two-dimensional model of the taut circular chain when the second imaging ray does not intersect with the lower surface of the chain pressure block.

[0055] Figure 7 Here is a flowchart for calculating the pitch of a circular link chain;

[0056] Figure 8 To acquire continuous raw images for the camera;

[0057] Figure 9 To Figure 8 Images that have undergone optical distortion correction;

[0058] Figure 10 To Figure 9 Images with perspective distortion correction;

[0059] Figure 11 To be Figure 10 The target detection image obtained by inputting the YOLOv8-seg model.

[0060] In the diagram: 1 is the L-shaped bracket, 2 is the tail of the scraper conveyor, 3 is the circular link chain, 4 is the camera, 5 is the middle plate of the tail, 6 is the optical center of the camera, 7 is the optical axis of the camera, 8 is the center of the tail sprocket, 9 is the first marker point, 10 is the second marker point, 11 is the third marker point, 12 is the fourth marker point, 13 is the calibration plane, 14 is the upper half of the blue mask area of ​​the flat ring, 15 is the lower half of the blue mask area of ​​the flat ring, 16 is the blue mask area of ​​the scraper, 17 is the imaging ray, 18 is the second imaging ray, 19 is the lower surface of the pressure block, 20 is the upper surface of the scraper, 21 is the centerline plane of the circular link chain, 22 is the centerline plane of the second circular link chain, 23 is the pitch circle of the tail sprocket, 24 is the upper surface of the second scraper, and 25 is the centerline plane of the circular link chain in the suspended state. Detailed Implementation

[0061] The present invention will be further illustrated by the following embodiments, but is not limited to the following embodiments.

[0062] Example 1:

[0063] Step 1: As Figures 1-2As shown, camera 4 is mounted on an L-shaped bracket 1 above the circular chain 3 in the tail section 2 of the scraper conveyor. The optical axis of the camera is as perpendicular as possible to the upper surface of the middle plate 5 in the tail section. Camera 4 is located at the center of the width direction of the middle plate 5 in the tail section. The field of view of camera 4 covers at least 8 circular chains 3 from the sprocket outlet. When the number of circular chains 3 between two adjacent scrapers is greater than 8, the number of circular chains 3 covered by the field of view of camera 4 is at least 1 more than the number of circular chains 3 between two adjacent scrapers.

[0064] Specific installation dimensions are as follows: Figure 3 As shown, the vertical distance from the camera's optical center 6 to the upper surface of the middle plate 5 at the tail of the camera is z.

[0065] Step Two: The process of obtaining basic parameters is as follows:

[0066] (1) Obtain the camera intrinsic parameter matrix and distortion coefficients

[0067] In a laboratory setting, multiple chessboard images were captured from different angles using a camera. The pixel positions of the chessboard corners were extracted from these images, and combined with the actual physical dimensions of the chessboard, the camera's intrinsic parameter matrix K and distortion coefficient d were calculated.

[0068] in, f x and f y c represents the equivalent focal length of the camera in the horizontal and vertical directions of the image, respectively. x and c y These represent the positions of the camera principal point in the image pixel coordinate system.

[0069] (2) Determine the image millimeter-to-pixel ratio under the working condition object distance.

[0070] In a laboratory setting, the field of view of the camera at the current object distance is determined based on the mounting distance *z* from the optical center of the camera to the upper surface of the center plate at the tail of the camera. Let W be the horizontal field of view width corresponding to this object distance. z The vertical field of view height is H. z The image has N horizontal pixels. x The vertical pixel count is N y Then, according to formula (1), the ratio of millimeter pixels s of the image in the horizontal and vertical directions at the object distance is calculated. x s y And calculate the average millimeter pixel ratio s of the image according to formula (2).

[0071] (1)

[0072] (2)

[0073] (3) Establish a calibration plane based on the captured images with markers and solve for the homography matrix.

[0074] like Figure 1 As shown, four crosshair markers are made on the upper surface of the middle plate 5 at the tail of the camera, within the field of view of camera 4. These are designated as marker 9, marker 10, marker 11, and marker 12. The arrangement of the four crosshair markers should meet the following requirements: the four crosshair markers form a rectangle in physical space; the center of the rectangle should coincide with the position corresponding to the main point of the camera as much as possible; the length direction of the rectangle should be consistent with the running direction of the circular chain 3; to ensure that the pitch of the circular chain 3 in each image can be calculated and that it contains at least one scraper, the effective length L of the rectangle in the length direction should meet the following requirements: when the number of circular chains 3 between two adjacent scrapers is less than or equal to 7, L should be greater than the standard pitch of 7 circular chains 3; when the number of circular chains 3 between two adjacent scrapers is greater than 7, L should be greater than the distance between two adjacent scrapers. The effective width of the rectangle in the width direction should be greater than twice the center distance of the chains. The rectangular area enclosed by the above four crosshair markers is defined as calibration plane 13.

[0075] Based on the camera intrinsic parameter matrix K and distortion coefficients obtained in the first step Optical distortion correction is performed on the image captured by the camera to obtain an optically distorted image. From this optically distorted image, the pixel coordinates of four crosshair markers are extracted and denoted as... The four crosshair markers are numbered according to their positional relationship in the optically distorted image: N1 (top left), N2 (bottom left), N3 (bottom right), and N4 (top right). Four target corner points corresponding to these four crosshair markers are then established in the calibration plane coordinate system, denoted as... Let the physical length of the calibration plane be L and the physical width be W. Then the four target corner points are taken as follows: , , , Source point N i With target point M i The plane homography relationship between them is satisfied, and the plane homography matrix H can be calculated according to formula (3).

[0076] (3)

[0077] In the formula: , , ;

[0078] Among them, h 11 h 12 h 13 h 21 h 22 h 23 h31 h 32 h 33 The matrix elements for calibrating the plane homography matrix H; is the scale factor; s is the average pixel-to-millimeter ratio of the image.

[0079] (4) Determine the image acquisition time interval

[0080] During the normal operation of the circular chain, the camera operates according to the image acquisition time interval. Continuous photography is performed. To ensure that no images of the circular chain are missed during continuous camera shooting, the image acquisition time interval needs to be adjusted. Reasonable settings should be made. To ensure that there are overlapping circular chain areas in two adjacent frames, the displacement of the circular chain between two adjacent shots should not exceed the physical length L of the calibration plane. The maximum allowable shooting interval Δt can be calculated according to formula (4) based on the maximum scraper chain speed v of the scraper conveyor and the physical length L of the calibration plane. max ,when This ensures that the circular chain is captured without any missed shots during camera photography.

[0081] (4)

[0082] (5) Accurately solve the pose relationship between the calibration plane and the camera optical center based on the captured images with markers.

[0083] Based on the camera intrinsic parameter matrix K obtained in the first step and the calibration plane homography matrix H obtained in the third step, construct the matrix... By decomposing matrix A, we can obtain the rotation matrix R of the calibration plane relative to the camera, the translation vector t, and the unit normal vector n of the calibration plane according to formula (5).

[0084] (5)

[0085] In the formula: , , , , , , ; UΣV is the approximate rotation matrix obtained from homography matrix decomposition. T To The result obtained by performing singular value decomposition is Σ, where Σ is the singular value matrix, and R = Udiag(1,1,det(UV)). T ))V T To The rotation matrix obtained after orthogonalization, where λ is the scale factor.

[0086] To convert the unit length in the calibration plane coordinate system to the actual physical length, the scale conversion factor 'a' can be calculated using the physical dimensions of the calibration plane according to formula (6). The scale conversion factor 'a' represents the proportional relationship between the unit length in the calibration plane coordinate system and the actual physical length. This step is used to determine the pose relationship between the calibration plane and the camera optical center, providing a basis for subsequently solving the position of the camera optical center in the two-dimensional model.

[0087] (6)

[0088] In the formula: , ;

[0089] Where W and L are the physical width and length of the calibration plane, respectively; W unit L unit The width and length of the calibration plane in the established planar coordinate system are given; M1, M2, and M4 are the target corner points in the calibration planar coordinate system established in the second step.

[0090] According to formula (7), the perpendicular distance dist between the camera optical center and the projection line of the calibration plane in the XZ plane can be calculated. The projection direction of the unit normal vector n of the calibration plane in the XZ plane is used.

[0091] The angle between it and the camera's optical axis can be calculated using formula (8). Using the camera principal point coordinates obtained in the first step... The homography transformation is used to map it onto the calibration plane coordinate system;

[0092] The coordinates of the intersection point E of the camera optical axis and the calibration plane in the calibration plane coordinate system can be calculated according to formula (9). .Depend on , , The pose relationship between the camera optical center and the calibration plane in the two-dimensional model can be determined.

[0093] (7)

[0094] (8)

[0095] (9)

[0096] In the formula: , is the unit normal vector of the calibration plane in the camera coordinate system; s is the average millimeter-to-pixel ratio of the image.

[0097] (6) Determine the precise position coordinates of the camera optical center in the tail of the scraper conveyor using the graphical method.

[0098] like Figure 3 As shown, in the two-dimensional model under the tension of the circular chain, the corresponding point of the camera optical center 6 in the two-dimensional model is P, the corresponding points of the first marker point 9 and the second marker point 10 in the two-dimensional model are B and C respectively, the corresponding point of the tail sprocket center 8 in the two-dimensional model is O, and point E is the intersection of the camera optical axis 7 and the calibration plane 13.

[0099] Establish an XZ plane coordinate system with the center 8 of the tail sprocket as the origin, and denote the center 8 of the tail sprocket as... Let the coordinates of the first marker point 9 and the second marker point 10 in the two-dimensional model be respectively... , Based on the distance measured on-site from the first marker point 9 to the center 8 of the tail sprocket... The distance from the first marker point 9 to the second marker point 10 is the physical length L of the marker plane. Since the first marker point 9 and the second marker point 10 are located on the upper surface of the middle plate 5 at the tail of the camera, the positions of points B and C in the XZ plane coordinate system can be determined, thus determining the position of the calibration plane 13 in the two-dimensional model. Let the coordinates of the intersection point E of the camera optical axis 7 and the calibration plane 13 in the XZ plane coordinate system be... Since the coordinates of the intersection point E of the camera optical axis 7 and the calibration plane 13 in the calibration plane coordinate system are... Therefore, in the two-dimensional model, the distance from the intersection point E of the camera optical axis 7 and the calibration plane 13 to the first marker point 9 is x. e Based on the fact that the intersection point E of the camera optical axis 7 and the calibration plane 13 is on the calibration plane 13, the position of point E in the XZ plane coordinate system can be determined, thereby determining the position of the intersection point E of the camera optical axis 7 and the calibration plane 13 in the two-dimensional model.

[0100] According to parameters , , The position coordinates of P in the XZ plane coordinate system can be determined. .

[0101] (7) Start the pitch calculation process

[0102] After completing the above six preliminary steps, the average millimeter-to-pixel ratio s, camera intrinsic parameter matrix K, and distortion coefficients of the image have been obtained. Calibration plane homography matrix H, image acquisition time interval Scale conversion factor The parameters also include the positional parameters of the crosshair calibration points B and C in the two-dimensional model and the camera optical center P. These parameters remain unchanged during system operation and serve as the basis for subsequent pitch calculations.

[0103] Based on this, the camera is started to acquire images continuously, and each image is processed sequentially to complete the calculation of the pitch of the circular chain.

[0104] Step 3: The pitch calculation process is as follows:

[0105] The specific process is as follows: Figure 7 As shown.

[0106] (1) Obtain the original image

[0107] Start the camera to continuously acquire raw images.

[0108] (2) Optical distortion correction

[0109] Using the camera intrinsic parameter matrix K and the distortion coefficient d, optical distortion correction is performed on the acquired original image to obtain an optically distorted image.

[0110] (3) Correction of perspective distortion

[0111] Using the calibrated plane homography matrix H, perspective distortion correction is performed on the optically distorted image to obtain a perspective-distorted image.

[0112] (4) Target detection and instance segmentation

[0113] like Figure 11 As shown, the perspective-dedistorted image is input into the instance segmentation model. The complete flat ring in the image is divided into upper and lower halves for detection and pixel-level segmentation, generating blue mask regions 14 and 15 for the upper and lower halves of the flat ring, respectively. At the same time, the scraper in the image is also detected and segmented at the pixel level, generating a blue mask region 16 for the scraper, and the corresponding binary masks are output to obtain the target detection image.

[0114] (5) Calculate the pixel coordinates of the centroid of the flat ring and the x-coordinates of the pixels at the left and right boundaries of the scraper.

[0115] In the target detection image, the binary masks corresponding to the upper half of the flat ring, the lower half of the flat ring, and the scraper are obtained respectively. The target detection image contains two circular chain chains, and the flat rings on each chain are independently numbered. Let the binary mask of the upper half of the k-th flat ring in the image be... The lower half of the binary mask is The binary mask for the i-th scraper from the left in the image is: ,in When the binary mask is set to 1, it represents a pixel. It belongs to the target area; otherwise, it does not belong to the area.

[0116] To calculate the centroid pixel coordinates of each mask region, firstly, the spatial moments m of each mask region are calculated according to formula (10). pq Then, according to formula (11), the centroid pixel coordinates of the upper half of the mask region of the k-th flat ring in the image are calculated. The centroid pixel coordinates of the lower half of the mask region of the k-th flat ring in the image are calculated according to formula (12). Then, according to formula (13), the centroid pixel coordinates of the upper and lower halves of the mask region of the kth flat ring in the image are synthesized to obtain the centroid pixel coordinates of the entire mask region of the kth flat ring in the image. Simultaneously, based on the binary mask of the i-th scraper in the image... Obtain the left and right horizontal coordinates of the i-th scraper in the horizontal direction of the image, denoted as ... , .

[0117] (10)

[0118] (11)

[0119] (12)

[0120] (13)

[0121] Where, m 00 m is the number of valid pixels within the mask region, i.e., the area of ​​the mask region; 10 m is the first spatial moment of the mask region about the vertical direction of the image; 01 Let the first-order spatial moment of the mask region be the image's horizontal direction; when When the value is zero, it indicates that the mask area has no valid pixels and is not included in the centroid calculation.

[0122] (6) Calculate the spatial position of the scraper and the circular chain in the two-dimensional model based on the pixel coordinates of the left and right boundaries of the scraper identified in the image.

[0123] Using the horizontal coordinates of the left and right boundaries of the first scraper in the image obtained by step (5) above , According to formula (14), it is converted into the horizontal coordinates in the calibration plane coordinate system, which are respectively , .

[0124] (14)

[0125] like Figure 3 As shown, in the two-dimensional model of the taut circular chain, the corresponding point of the camera optical center 6 in the two-dimensional model is P, 17 is the imaging ray, 18 is the second imaging ray, the UV segment is the arc segment of the lower surface 19 of the pressure block, and the V point extends to the right to form the straight line segment of the lower surface 19 of the pressure block.

[0126] In the diagram, the first marker point 9 and the second marker point 10 correspond to points B and C in the two-dimensional model, respectively, and the center point 8 of the tail sprocket corresponds to point O in the two-dimensional model. Points S1, S2, S3, and S4 are the four endpoints of the scraper, and points G and F are the connection points between the inner flat ring of the scraper and the adjacent vertical ring. Points G and F are the midpoints of line segments S1S3 and S2S4, respectively. Points T1, T2, T3, and T4 are the four endpoints of the second scraper, and point J is the connection point between the inner flat ring of the second scraper and the adjacent vertical ring. Point J is the midpoint of line segments T1T3. Point Q is the tangent point between the plane 21 of the circular chain centerline and the pitch circle 23 of the tail sprocket.

[0127] Establish an XZ plane coordinate system with the center 8 of the tail sprocket as the origin, and denote the center 8 of the tail sprocket as... Record the coordinates of O, B, C, P, U, and V in this coordinate system.

[0128] Find the horizontal coordinates of the left and right boundaries of the first scraper from the left in the image in the calibration plane coordinate system. , Transform to the XZ plane coordinate system of the two-dimensional model, and denote the corresponding points in the two-dimensional model as follows: , The distance from point A1 to the first marker point 9 is... ,point The distance to the first marker point 9 is According to point A1, On calibration plane 13, point A1 can be determined. Coordinates in a two-dimensional model.

[0129] Determine the x-coordinate of the right boundary of the first scraper from the left in the image. Is it equal to the x-coordinate of the right boundary of the image? ,like = Calculate according to step 1) below, if ≠ Then calculate according to step 2) below.

[0130] 1) When = At that time, the left boundary of the upper surface 20 of the scraper will definitely be fully displayed in the image. For example... Figure 4 As shown, at this point, take point A1, and connect the camera's optical center point P with point A1 to obtain the imaging ray 17 corresponding to the left boundary of the upper surface 20 of the scraper. The intersection of this imaging ray 17 and the lower surface 19 of the chain pressing block is the actual position of the left boundary of the upper surface 20 of the scraper in the two-dimensional model, denoted as . .

[0131] With the circular chain taut, the upper surface 20 of the scraper is tangent to the arc segment UV of the lower surface 19 of the pressure block, and the scraper has fixed structural dimensions. Let the width of the scraper in the two-dimensional model be a and the height be b. The actual position of the left boundary of the upper surface 20 of the scraper is then determined. Then, the positions of the other three endpoints S2, S3, and S4 of the scraper in the two-dimensional model are determined. From this, the coordinates of the connection points G and F between the inner flat ring and the adjacent vertical ring of the scraper can be determined. Since the centerline plane 21 of the circular chain is tangent to the pitch circle 23 of the tail sprocket, point Q is its point of tangency. The coordinates of point Q can be determined using the coordinates of point G.

[0132] 2) When ≠ At this time, the right boundary of the upper surface 20 of the scraper will be fully displayed in the image. At this point, the point is selected. Connect the camera's optical center point P with point P. Obtain the second imaging ray 18 corresponding to the right boundary of the upper surface 20 of the scraper, and determine whether the second imaging ray 18 intersects with the lower surface 19 of the chain pressing block. If there is an intersection, calculate according to step ①; otherwise, calculate according to step ②.

[0133] ①For example Figure 5 As shown, there is an intersection point, which is the actual position of the right boundary of the upper surface 20 of the scraper in the two-dimensional model, denoted as point. Similarly, the actual position of the right boundary of the upper surface 20 on the scraper is determined. Then, the positions of the other three endpoints S1, S3, and S4 of the scraper in the two-dimensional model are determined, and the coordinates of the connection points G and F between the inner flat ring and the adjacent vertical ring of the scraper can be determined from this. Furthermore, the coordinates of the tangent point Q between the center line plane 21 of the circular chain and the pitch circle 23 of the tail sprocket can be determined from the coordinates of point G.

[0134] Based on this, according to the fixed length constraint connecting the current scraper and the second scraper through the center line plane 22 of the second circular chain, let the length of the center line plane 22 of the second circular chain be e. Then, based on the position of point V, determine whether the right boundary of the upper surface 24 of the second scraper is located on the arc segment UV of the lower surface 19 of the pressure block or on the straight line segment after point V in the two-dimensional model.

[0135] When the right boundary of the upper surface 24 of the second scraper is located on the arc segment UV of the lower surface 19 of the chain pressure block, the upper surface 24 of the second scraper is tangent to the arc segment when the circular chain is taut. The coordinates of the four endpoints T1, T2, T3, and T4 of the second scraper in the two-dimensional model, as well as the coordinates of the connection point J between the inner flat ring and the adjacent vertical ring of the second scraper, can be determined.

[0136] When the right boundary of the upper surface 24 of the second scraper is located on the straight line segment after point V on the lower surface 19 of the pressure block, the upper surface 24 of the second scraper coincides with this straight line segment when the circular link chain is taut. The coordinates of the four endpoints T1, T2, T3, T4 and the connection point J of the second scraper in the two-dimensional model can be determined.

[0137] ②For example Figure 6 As shown, there are no intersection points. The actual position of the right boundary of the upper surface 20 of the scraper in the two-dimensional model is located on the second imaging ray 18, and the center line plane 21 of the circular chain and the center line plane 22 of the second circular chain are collinear, that is, points Q, G, F, and J are collinear. The right boundary of the upper surface 24 of the second scraper is located on the arc segment UV of the lower surface 19 of the pressure block in the two-dimensional model. The scraper and the second scraper are connected by the center line plane 22 of the second circular chain, and its length is e.

[0138] With the circular chain taut, the upper surface 24 of the second scraper is tangent to the arc segment. The coordinates of the four endpoints S1, S2, S3, and S4 of the scraper in the two-dimensional model can be determined. From this, the coordinates of the connection points G and F between the inner flat ring and the adjacent vertical ring of the scraper, as well as the coordinates of the four endpoints T1, T2, T3, and T4 of the second scraper in the two-dimensional model, can be determined. From this, the coordinates of the connection point J between the inner flat ring and the adjacent vertical ring of the second scraper, and the coordinates of the tangent point Q between the centerline plane 21 of the circular chain and the pitch circle 23 of the tail sprocket, can be determined.

[0139] (7) Convert the pixel coordinates of the centroid of the flat ring into the physical coordinates of the calibration plane.

[0140] like Figure 5 As shown, perpendicular lines are drawn from points G and F to the calibration plane 13, with feet at points Z0 and Z1 respectively. Let the distance from point Z0 to point B be denoted as . The distance from point Z1 to point B is Then we have: , Let D0 be the distance from point G to point Z0, and D1 be the distance from point F to point Z1. Then we have: , Let the straight line be... The angle between the line and line BC is The angle between line FJ and line BC is Then we have: , .

[0141] The center plane 21 of the circular link chain can be considered as rotating around the rotation center Z0 from the calibration plane 13. After translating D0, the second circular chain centerline plane 22 can be considered as rotating from the calibration plane 13 around the rotation center Z1. And obtained by translating D1.

[0142] The pixel coordinates of the k-th centroid of the flat ring in the image obtained in step (5) above. According to formula (15), it is converted into physical coordinates on the calibration plane. .

[0143] (15)

[0144] In the formula: , , , , , , .

[0145] Where, when the pixel x-coordinate u of the centroid of the k-th flat ring in the image k smaller than the left boundary pixel value of the scraper hour, , , When the pixel x-coordinate u of the centroid of the k-th flat ring in the image k The pixel value greater than the right boundary of the scraper hour, , , d is the diameter of the circular chain.

[0146] (8) Calculate the pitch of the circular link chain under tension.

[0147] Using the physical coordinates of the kth and (k+1)th centroids of the horizontal rings obtained in step (7) above on the calibration plane , The pitch P of the circular link chain under tension is calculated according to formula (16). k When there is no scraper between two adjacent flat rings, the distance between the centers of gravity of the two flat rings is twice the pitch of the circular chain, and n=2; when there is a scraper between two adjacent flat rings, the distance between the centers of gravity of the two flat rings is four times the pitch of the circular chain, and n=4.

[0148] (16)

[0149] (9) Calculate the horizontal link distance correction factor when considering chain suspension.

[0150] like Figure 4As shown, due to the influence of tension changes during operation, the circular chain is not necessarily in a taut state and may also be in a suspended state. The figure shows the centerline plane 25 of the circular chain in the suspended state. When the circular chain is in a suspended state, its spatial position shifts relative to the taut state, and its imaging effect can be equivalent to an increase in the object distance from the corresponding circular chain to the camera's optical center 6. According to the perspective imaging principle of near objects appearing larger and far objects appearing smaller, when the object distance increases, the projected size of the flat ring in the image will decrease, thus making the pixel distance between the centers of gravity of the upper and lower halves of the flat ring smaller than the corresponding pixel distance in the taut state. During the operation of the circular chain under load, its pitch direction deformation is significant, while the deformation in the center of gravity direction of the upper and lower halves of the circular chain is smaller. Therefore, the distance between the centers of gravity of the upper and lower halves of the flat ring is used as a reference quantity to characterize the change in object distance.

[0151] Therefore, the theoretical pixel distance between the centroids of the upper and lower halves of the k-th flat ring in the image under tension is first calculated according to formula (17). Combined with the pixel distance between the centroids of the upper and lower halves of the k-th flat ring actually measured in the image, The k-th horizontal distance correction coefficient r is constructed according to formula (18). k And use this coefficient to determine the pitch P of the circular chain obtained in step (8) under tension. k Compensation corrections are made to improve the accuracy of the circular link chain pitch measurement results.

[0152] (17)

[0153] (18)

[0154] In the formula: , , , , , , , , , , , ; ;

[0155] Among them, P 1k P 2k q represents the physical coordinates of the centroids of the upper and lower halves of the k-th flat ring in the image on the calibration plane; 1k q 2k These are the pixel coordinates of the centroids of the upper and lower halves of the k-th flat ring, respectively; when the x-coordinate of the centroid of the k-th flat ring in the image is u... k smaller than the left boundary pixel value of the scraper hour, , , When the pixel x-coordinate u of the centroid of the k-th flat ring in the image k The pixel value greater than the right boundary of the scraper hour, , , ; d is the diameter of the circular chain; s y This is the standard physical distance between the centers of gravity of the upper and lower halves of the flat ring.

[0156] (10) Calculate the corrected pitch of the circular chain.

[0157] The pitch P of the circular chain obtained in step (8) under tension is k The distance between the centroids of the kth and (k+1)th horizontal rings is obtained, and therefore subject to the correction coefficients for the distances of the kth and (k+1)th horizontal rings. , The combined effect of these factors. Therefore, the tension coefficients of the kth and (k+1)th flat rings are averaged, and the pitch P of the circular chain under tension is calculated according to formula (19). k Perform compensation and correction, and calculate the final circular chain pitch P. k final :

[0158] (19)

[0159] (11) Output the calculated pitch of the circular chain.

[0160] The specific analysis and calculation process of this embodiment is as follows:

[0161] The equipment and parameter requirements used in this embodiment are as follows:

[0162] (1) The main technical parameters of the scraper conveyor are: Model: SGZ1000 / 1400; Scraper chain speed: 0-1.3m / s; Chain type: mining high-strength round link chain; Round link chain specifications: 42×146mm; Groove width: 1000mm; Chain center distance: 220±1.0mm; Scraper spacing: 1168mm; Sprocket pitch circle diameter: 750mm; Number of adjacent scraper round link chains: 7; Scraper height: 115mm; Scraper width: 135mm; Distance between the center of gravity of the upper and lower halves of the flat ring: 86.824mm.

[0163] (2) The main specifications of the industrial camera are as follows: Model: DS-2CD2686FWDA2-IZS; Resolution: 3840(H)×2160(V); Frame rate: 25fps; Exposure time: 1us~1 / 3s; Focal length: 2.7~12mm; Horizontal field of view: 108.1°~45.6°; Vertical field of view: 58.4°~25.7°; Diagonal field of view: 127.4°~52.2°; Image acquisition parameter requirements: Exposure time: 100us.

[0164] (3) The main specifications of the computer are as follows: Processor: Intel Xeon 8370; Memory: 32G DDR4 3200MHZ; Hard disk: 480G SSD + 4T mechanical hard disk; Graphics card: RTX A4000-16G.

[0165] The specific steps for taking the measurement are as follows:

[0166] Step 1: The camera is mounted on the tail support of the scraper conveyor, approximately at the center of the width of the middle plate at the tail. For example... Figure 3 As shown: The vertical distance from the camera optical center 6 to the upper surface of the middle plate 5 at the tail is 1260mm.

[0167] The second step, obtaining the basic parameters, is as follows:

[0168] (I): Obtaining the camera intrinsic parameter matrix :

[0169] [[2.74648667e+03, 0.00000000e+00, 1.89401746e+03],

[0170] [0.00000000e+00, 2.74572638e+03, 1.09356485e+03],

[0171] [0.00000000e+00, 0.00000000e+00, 1.00000000e+00]],

[0172] Distortion coefficient d: [-3.50885894e-01, 1.80144321e-01, 2.83153426e-04, 4.77632550e-04, -6.48590241e-02].

[0173] (II): The camera is in At object distance, the corresponding horizontal field of view width Vertical field of view height Image horizontal pixel count Vertical pixel count The horizontal pixel ratio in millimeters is obtained according to formula (1). The image's vertical millimeter-pixel ratio The average millimeter-to-pixel ratio of the image is obtained according to formula (2). .

[0174] (iii): Physical length of the calibration plane Physical width The pixel coordinates of four crosshair markers are extracted from the optically distorted image. Cross marker Cross marker Cross marker The pixel coordinates are (751, 370), (769, 1350), (2459, 1255), and (2459, 402). Four corresponding corner points are then constructed in the calibration plane coordinate system. , corner point , corner point , corner point The coordinates are (0, 0), (0, 1278), (2391, 1278), and (2391, 0). The homography matrix H of the calibration plane is obtained according to formula (3):

[0175] [[1.14794578e+00, -2.10847183e-02, -8.54305932e+02],

[0176] [-2.27992525e-02, 1.21691010e+00, -4.33134499e+02],

[0177] [-7.18614673e-05, -8.81836818e-06, 1.00000000e+00]]

[0178] (iv): The maximum scraper chain speed of the scraper conveyor and calibration plane physical length The maximum allowable photo interval can be obtained according to formula (4). Determine the image acquisition time interval ;

[0179] (v): Constructed matrix for:

[0180] [[0.000406, 0.000003, -0.4959],

[0181] [-0.000007, 0.000412, -0.250531],

[0182] [0.000067, 0.000009, 1.060705]]

[0183] The rotation matrix of the calibration plane relative to the camera is obtained according to formula (5). for:

[0184] [[0.986689, 0.011415, -0.162219],

[0185] [-0.014822, 0.999694, -0.019807],

[0186] [0.161943, 0.021948, 0.986556]]

[0187] Translation vector for:

[0188] [-1204.30296029, -608.41906158, 2575.94117433]

[0189] Unit normal vector of the calibration plane for:

[0190] [-0.16221868, -0.0198075, 0.98655601]

[0191] The size conversion factor is obtained according to formula (6). .

[0192] (vi): Calculate the perpendicular distance between the optical center of the camera and the projection line of the calibration plane in the XZ plane according to formula (7). .

[0193] The unit normal vector of the calibration plane is obtained according to formula (8). The angle between the projection direction in the XZ plane and the camera optical axis .

[0194] The intersection point of the camera's optical axis and the calibration plane is obtained using formula (9). Coordinates in the calibration plane coordinate system .

[0195] (vii): such as Figure 3 As shown, the distance from the first marker point 9 to the center 8 of the tail sprocket is... , , , , .

[0196] The third step, pitch calculation, is as follows:

[0197] Step 1: After mounting and securing the camera and configuring the acquisition parameters, start the camera to continuously acquire raw images, such as... Figure 8 As shown.

[0198] Step 2: Using the camera intrinsic parameter matrix K and distortion coefficients Optical distortion correction is performed on the acquired original image to obtain an optically distorted image, such as... Figure 9 As shown.

[0199] Step 3: Using the calibrated planar homography matrix H, perform perspective distortion correction on the optically distorted image to obtain the perspective-distorted image, such as... Figure 10 As shown.

[0200] Step 4: Input the perspective-dedistorted image into the YOLOv8-seg model to obtain the target detection image, such as... Figure 11 As shown.

[0201] Step 5: In the target detection image, such as Figure 11 According to formulas (11) and (12),

[0202] The centroid pixel coordinates of the upper half of the mask region of the first flat loop on the chain ;

[0203] The centroid pixel coordinates of the lower half of the mask region of the first flat loop on the chain ;

[0204] The centroid pixel coordinates of the upper half of the mask region of the second flat ring on the chain ;

[0205] The centroid pixel coordinates of the lower half of the mask region of the second flat loop on the upper chain .

[0206] The centroid pixel coordinates of the upper half of the mask region of the first flat loop in the next chain ;

[0207] The centroid pixel coordinates of the lower half of the mask region of the first flat loop in the next chain ;

[0208] The centroid pixel coordinates of the upper half of the mask region of the second flat loop in the next chain ;

[0209] The centroid pixel coordinates of the lower half of the mask region of the second flat loop in the next chain .

[0210] According to formula (13),

[0211] The centroid pixel coordinates of the first flat ring global mask region on the chain ;

[0212] The centroid pixel coordinates of the second flat ring overall mask region on the chain ;

[0213] The centroid pixel coordinates of the first flat ring global mask region in the next chain ;

[0214] The centroid pixel coordinates of the second flat ring overall mask region in the next chain .

[0215] Obtain the x-coordinate of the left boundary of the scraper in the horizontal direction of the image. ;

[0216] Obtain the x-coordinate of the right boundary of the scraper in the horizontal direction of the image. .

[0217] Step 6: According to formula (14), the horizontal coordinates of the left and right boundaries of the scraper in the horizontal direction of the image are calculated. , Converted to the horizontal coordinates in the calibration plane coordinate system, respectively , .

[0218] The left and right boundaries of the scraper are located in the horizontal coordinate system of the calibration plane. , Transformed into the XZ plane coordinate system of the two-dimensional model, respectively , .

[0219] The right boundary x-coordinate of the scraper in the horizontal direction of the image The right boundary x-coordinate of the unequal graph Therefore, proceed to step three (6), 2). Select a point. Point P and point The line connecting the two points intersects with the arc segment UV, so step three (6), 2)① is executed. This intersection point is the actual position of the right boundary of the upper surface of the scraper in the two-dimensional model. . , , .

[0220] The coordinates of point G are (1276.7, -116.1), the coordinates of point F are (1408.3, -146.3), and the coordinates of point Q are (141.7, 347.2).

[0221] Step 7: Click The coordinates are (1256.7, -171.2). , , According to formula (15), the image is obtained as follows:

[0222] The physical coordinates of the centroid of the first flat loop on the calibration plane ;

[0223] The physical coordinates of the centroid of the second flat loop on the calibration plane ;

[0224] The physical coordinates of the centroid of the first flat loop of the next chain on the calibration plane ;

[0225] The physical coordinates of the centroid of the second flat loop of the lower chain on the calibration plane .

[0226] Step 8: There is no scraper between two adjacent flat rings, so n=2. The pitch of the upper chain ring in the image under tension is obtained according to formula (16). Pitch of the lower chain link in a taut state ;

[0227] Step 9: Calculate the theoretical pixel distance between the centers of gravity of the upper and lower halves of the first flat loop of the upper chain in the image when it is under tension, according to formula (17). ;

[0228] The theoretical pixel distance between the centers of gravity of the upper and lower halves of the second flat loop in the upper winding when it is taut ;

[0229] The theoretical pixel distance between the centers of gravity of the upper and lower halves of the first flat loop of the lower chain when it is taut ;

[0230] The theoretical pixel distance between the centers of gravity of the upper and lower halves of the second flat loop in the lower chain when it is taut .

[0231] Pixel distance between the centroids of the upper and lower halves of the first flat loop on the chain ;

[0232] The pixel distance between the centroids of the upper and lower halves of the second flat loop on the upper chain ;

[0233] Pixel distance between the centroids of the upper and lower halves of the first flat loop in the next chain ;

[0234] Pixel distance between the centroids of the upper and lower halves of the second flat loop in the lower chain .

[0235] The tightness coefficient of the first flat loop in the upper chain in the image is obtained according to formula (18). ;

[0236] Tightness coefficient of the second flat loop on the upper chain ;

[0237] Tightness coefficient of the first flat ring of the lower chain ;

[0238] Tightness coefficient of the second flat ring of the lower chain .

[0239] Step 10: Calculate the final loop pitch of the upper chain in the image using formula (19). ;

[0240] The final link pitch of the lower chain .

Claims

1. A method for measuring the pitch of a circular link chain in a mining scraper conveyor, characterized in that, Includes the following steps: Step 1: Equipment Installation and Debugging The camera is installed above the middle circular chain at the tail of the scraper conveyor, with the camera's optical axis perpendicular to the bottom arc of the tail groove. The camera's field of view covers 8 to 10 circular chains at the sprocket outlet. The camera was calibrated to obtain the average millimeter-to-pixel ratio s, camera intrinsic parameter matrix K, distortion coefficient d, calibration plane homography matrix H, and image acquisition time interval. The scale conversion factor a, and the position parameters of the crosshair calibration points B and C and the camera optical center P in the two-dimensional model; Step 2: Image Processing and Pitch Calculation (1) Acquire raw images: Start the camera to acquire raw images continuously; (2) Optical distortion correction: Using the camera intrinsic parameter matrix K and distortion coefficient d, optical distortion correction is performed on the acquired original image to obtain an optically distorted image; (3) Perspective distortion correction: Using the calibrated plane homography matrix H, perspective distortion correction is performed on the optically distorted image to obtain the perspective-distorted image; (4) Target detection and instance segmentation: Input the perspective-dedistorted image into the instance segmentation model, and detect and segment the complete flat ring in the image into upper and lower halves, respectively, generating blue mask regions for the upper half and lower half of the flat ring; at the same time, detect and segment the scraper in the image, generate blue mask regions for the scraper, and output the corresponding binary masks to obtain the target detection image; (5) Calculate the pixel coordinates of the centroid of the flat ring and the horizontal coordinates of the left and right boundaries of the scraper: In the target detection image, obtain the binary mask corresponding to the upper half of the flat ring, the lower half of the flat ring, and the scraper respectively. (6) Calculate the spatial position of the scraper and the circular chain in the two-dimensional model based on the pixel coordinates of the left and right boundaries of the scraper identified in the image: using the left and right boundary x-coordinates of the first scraper in the image obtained in step (5) above. , Convert them to the horizontal coordinates in the calibration plane coordinate system, respectively. , ; (7) Convert the pixel coordinates of the centroid of the flat ring into the physical coordinates of the calibration plane; (8) Calculate the pitch of the circular link chain under tension: Using the physical coordinates of the kth and (k+1)th centroids of the horizontal rings obtained in step (7) above on the calibration plane , Calculate the pitch P of the circular link chain under tension. k ; (9) Calculate the horizontal link distance correction factor r when considering chain suspension. k r k+1 ; (10) Calculate the compensated and corrected pitch of the circular chain, and the final pitch of the circular chain. for: 。 2. The online circular link chain pitch measurement method for a mining scraper conveyor according to claim 1, characterized in that, In step two (1), four cross markers are made on the upper surface of the middle plate of the tail section within the camera's field of view. These are the first marker, the second marker, the third marker, and the fourth marker. The arrangement of the four cross markers should meet the following requirements: the four cross markers form a rectangle in physical space; the center of the rectangle formed by the four cross markers should coincide with the position corresponding to the main point of the camera as much as possible; and the length direction of the rectangle should be consistent with the running direction of the circular chain.

3. The online circular link chain pitch measurement method for a mining scraper conveyor according to claim 2, characterized in that, In step two (5), the target detection image contains two circular chains, and the flat loops on each circular chain are independently numbered; let the upper half of the binary mask of the k-th flat loop in the image be... The lower half of the binary mask is The binary mask for the i-th scraper from the left in the image is: ,in When the binary mask value is 1, it represents a pixel. It belongs to the target area; otherwise, it does not belong to the area.

4. The online circular link chain pitch measurement method for a mining scraper conveyor according to claim 3, characterized in that, In step two (6), in the two-dimensional model under the tension of the circular chain, the corresponding point of the camera optical center in the two-dimensional model is P, the UV segment is the arc segment of the lower surface of the pressure block, and the V point extends to the right to form the straight line segment of the lower surface of the pressure block; the corresponding points of the first and second marker points in the two-dimensional model are B and C, respectively, and the corresponding point of the tail sprocket center in the two-dimensional model is O; points S1, S2, S3, and S4 are the four endpoints of the scraper, and points G and F are the connection points between the inner flat ring of the scraper and the adjacent vertical ring; points G and F are the midpoints of line segments S1S3 and S2S4, respectively; points T1, T2, T3, and T4 are the four endpoints of the second scraper, and point J is the connection point between the inner flat ring of the second scraper and the adjacent vertical ring; among them, point J is the midpoint of line segment T1T3; point Q is the tangent point between the plane of the circular chain centerline and the pitch circle of the tail sprocket; Establish an XZ plane coordinate system with the center of the tail sprocket as the origin, and denote the center of the tail sprocket as... Record the coordinates of O, B, C, P, U, and V in this coordinate system. Find the horizontal coordinates of the left and right boundaries of the first scraper from the left in the image in the calibration plane coordinate system. , Transform to the XZ plane coordinate system of the two-dimensional model, and denote the corresponding points in the two-dimensional model as follows: , Where, the distance from point A1 to the first marker is ,point The distance to the first marker is According to point A1, On the calibration plane, point A1 can be determined. Coordinates in a two-dimensional model; Determine the x-coordinate of the right boundary of the first scraper from the left in the image. Is it equal to the x-coordinate of the right boundary of the image? ,like Calculate according to step 1) below, if Then calculate according to step 2) below: 1) When At that time, the left boundary of the upper surface of the scraper will be fully displayed in the image; take point A1, connect the camera's optical center point P with point A1 to obtain the imaging ray corresponding to the left boundary of the upper surface of the scraper, and the intersection of this imaging ray with the lower surface of the chain pressing block is the actual position of the left boundary of the upper surface of the scraper in the two-dimensional model, denoted as . ; With the circular chain taut, the upper surface of the scraper is tangent to the arc segment UV of the lower surface of the pressure block, and the scraper has fixed structural dimensions. Let the width of the scraper in the two-dimensional model be a and the height be b; determine the actual position of the left boundary of the upper surface of the scraper. Then, the positions of the other three endpoints S2, S3, and S4 of the scraper in the two-dimensional model are determined; and from this, the coordinates of the connection points G and F between the inner flat ring and the adjacent vertical ring of the scraper can be determined; since the plane of the center line of the circular chain is tangent to the pitch circle of the tail sprocket, point Q is its point of tangency, and the coordinates of point Q can be determined by the coordinates of point G. 2) When At that time, the right boundary of the upper surface of the scraper will definitely be fully displayed in the image; point selection Connect the camera's optical center point P with point P. Obtain the second imaging ray corresponding to the right boundary of the upper surface of the scraper, and determine whether the second imaging ray intersects with the lower surface of the chain pressing block. If there is an intersection, calculate according to step ①; if there is no intersection, calculate according to step ②. ① There is an intersection point, which is the actual position of the right boundary of the upper surface of the scraper in the two-dimensional model, denoted as point. ; Determine the actual position of the right boundary of the upper surface of the scraper Then, the positions of the other three endpoints S1, S3, and S4 of the scraper in the two-dimensional model are determined, and the coordinates of the connection points G and F between the inner flat ring and the adjacent vertical ring can be determined from this. Furthermore, the coordinates of the point Q, the point of tangency between the center line plane of the circular chain and the pitch circle of the tail sprocket, can be determined through the coordinates of point G. Based on this, according to the fixed length constraint of the connection between the scraper and the second scraper through the center line plane of the second circular chain, let the length of the center line plane of the second circular chain be e. Then, by using the position of point V, determine whether the right boundary of the upper surface of the second scraper is located on the arc segment UV of the lower surface of the pressure block or on the straight line segment after point V in the two-dimensional model: When the right boundary of the upper surface of the second scraper is located on the arc segment UV of the lower surface of the pressure block, combined with the fact that the upper surface of the second scraper is tangent to the arc segment when the circular chain is taut, the coordinates of the four endpoints T1, T2, T3, and T4 of the second scraper in the two-dimensional model, as well as the coordinates of the connection point J between the inner flat ring and the adjacent vertical ring of the second scraper, can be determined; When the right boundary of the upper surface of the second scraper is located on the straight line segment after point V of the lower surface of the pressure block, combined with the fact that the upper surface of the second scraper coincides with the straight line segment when the circular chain is taut, the coordinates of the four endpoints T1, T2, T3, and T4 of the second scraper and the connection point J in the two-dimensional model can be determined. ② There is no intersection point. The actual position of the right boundary of the upper surface of the scraper in the two-dimensional model is on the second imaging ray. The plane of the center line of the circular chain is collinear with the plane of the center line of the second circular chain, that is, points Q, G, F and J are collinear. The right boundary of the upper surface of the second scraper is located on the arc segment UV of the lower surface of the pressure block in the two-dimensional model. The scraper and the second scraper are connected by the plane of the center line of the second circular chain, and its length is e. When the circular chain is taut, and the upper surface of the second scraper is tangent to the arc segment, the coordinates of the four endpoints S1, S2, S3, and S4 of the scraper in the two-dimensional model can be determined. From this, the coordinates of the connection points G and F between the inner flat ring and the adjacent vertical ring of the scraper, as well as the coordinates of the four endpoints T1, T2, T3, and T4 of the second scraper in the two-dimensional model, can be determined. From this, the coordinates of the connection point J between the inner flat ring and the adjacent vertical ring of the second scraper, as well as the coordinates of the tangent point Q between the centerline plane of the circular chain and the pitch circle of the tail sprocket can be determined.

5. The online circular link chain pitch measurement method for a mining scraper conveyor according to claim 4, characterized in that, In step two (7), in the two-dimensional model of the taut circular chain, perpendicular lines are drawn from points G and F to line segment BC, with the feet of the perpendiculars at points Z0 and Z1, respectively; the distance from point Z0 to point B is denoted as . The distance from point Z1 to point B is Then we have: , Let D0 be the distance from point G to point Z0, and D1 be the distance from point F to point Z1. Then we have: , Let the angle between line QG and line BC be . The angle between line FJ and line BC is Then we have: , ;The pixel coordinates of the k-th centroid of the flat ring in the image obtained in step two (5) above. Convert to physical coordinates on the calibration plane using the following formula. ; ； In the formula: , , , , , , ; Where, when the pixel x-coordinate u of the centroid of the k-th flat ring in the image k smaller than the left boundary pixel value of the scraper hour, , , When the pixel x-coordinate u of the centroid of the k-th flat ring in the image k The pixel value greater than the right boundary of the scraper hour, , , d is the diameter of the circular chain.

6. The online circular link chain pitch measurement method for a mining scraper conveyor according to claim 5, characterized in that, In step two (8), the pitch P of the circular chain under tension is calculated according to the following formula. k : ； When there is no scraper between two adjacent flat rings, the distance between the centers of gravity of the two flat rings is twice the pitch of the circular chain, and n=2; when there is a scraper between two adjacent flat rings, the distance between the centers of gravity of the two flat rings is four times the pitch of the circular chain, and n=4.

7. The online circular link chain pitch measurement method for a mining scraper conveyor according to claim 6, characterized in that, In step two (9), the theoretical pixel distance between the centers of gravity of the upper and lower halves of the k-th flat ring in the image under tension is calculated according to the following formula. Combined with the pixel distance between the centroids of the upper and lower halves of the k-th flat ring actually measured in the image, The k-th parallel distance correction coefficient r is obtained. k And use this coefficient to determine the pitch P of the circular chain obtained in step two (8) under tension. k Compensation and correction are performed to improve the accuracy of the circular link chain pitch measurement results: ； ； In the formula: , , , , , , , , , , , ; ; Among them, P 1k P 2k q represents the physical coordinates of the centroids of the upper and lower halves of the k-th flat ring in the image on the calibration plane; 1k q 2k These are the pixel coordinates of the centroids of the upper and lower halves of the k-th flat ring, respectively; when the x-coordinate of the centroid of the k-th flat ring in the image is u... k smaller than the left boundary pixel value of the scraper hour, , , When the pixel x-coordinate u of the centroid of the k-th flat ring in the image k The pixel value greater than the right boundary of the scraper hour, , , ; d is the diameter of the circular chain; s y This is the standard physical distance between the centers of gravity of the upper and lower halves of the flat ring.

8. The online circular link chain pitch measurement method for a mining scraper conveyor according to claim 1, characterized in that, In step two (10), the pitch P of the circular chain obtained in step two (8) under tension is... k The distance between the centroids of the kth and (k+1)th horizontal rings is obtained, therefore it is subject to the correction coefficient r of the distance between the kth and (k+1)th horizontal rings. k r k+1 The combined effects; the tension coefficients of the kth and (k+1)th flat rings are averaged, and the pitch P of the circular chain under tension is calculated according to the formula. k Perform compensation and correction, and calculate the final pitch of the circular chain. for: 。

Images