Belt offset detection method, device and equipment and storage medium

Abstract

The invention relates to the technical field of image processing, in particular to a belt offset detection method, device and equipment and a storage medium, and the method comprises the steps: determining a target elliptical point set corresponding to a carrier roller in a reference scale belt conveyor image, and determining a carrier roller edge midpoint based on the target elliptical point set; determining a belt edge straight line based on the reference scale belt conveyor image, and calculating a projection point set of the belt edge straight line on the target elliptical point set; determining a belt edge midpoint based on the projection point set; and determining an offset detection result based on the midpoint of the edge of the carrier roller and the midpoint of the edge of the belt. The belt deviation detection device is convenient to improve the convenience of belt deviation detection.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a belt deviation detection method, device, equipment and storage medium. Background Art

[0002] Coal enterprises generally use belt conveyors to transport coal. Due to the wear of the belt or uneven feeding position on the belt conveyor, the belt will shift. The shift of the belt will lead to adverse effects such as the shift of the feeding direction. Therefore, during the feeding process of the belt conveyor, the belt needs to be detected for shift.

[0003] Currently, one way to detect belt deviation is to set physical sensors on both sides of the belt conveyor, such as touch sensors, micro switches, photoelectric switches, etc. Once the belt deviates from the preset track, the physical sensor will be triggered to send a corresponding signal to alert the staff.

[0004] However, physical sensors are more sensitive to environmental changes, such as temperature, humidity, dust, etc. Environmental changes will affect the detection of physical sensors. It can be seen that the existing technology is not convenient for detecting belt deviation. Summary of the Invention

[0005] In order to improve the convenience of detecting belt deviation, the present application provides a belt deviation detection method, device, equipment and storage medium.

[0006] In a first aspect, the present application provides a belt deviation detection method, comprising:

[0007] Determine a target ellipse point set corresponding to the idler in the reference scale belt conveyor image, and determine the midpoint of the idler edge based on the target ellipse point set;

[0008] Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set;

[0009] determining a belt edge midpoint based on the projected point set;

[0010] An offset detection result is determined based on the midpoint of the roller edge and the midpoint of the belt edge.

[0011] In a second aspect, the present application provides a belt deviation detection device, comprising:

[0012] A roller midpoint determination module is used to determine a target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determine the roller edge midpoint based on the target ellipse point set;

[0013] A projection point set generating module is used to determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set;

[0014] a belt midpoint determination module, configured to determine a belt edge midpoint based on the projection point set;

[0015] The detection result generating module is used to determine the offset detection result based on the midpoint of the edge of the roller and the midpoint of the edge of the belt.

[0016] In a third aspect, the present application provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above method when executing the computer program.

[0017] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the steps in the above-mentioned method when executed by a processor.

[0018] In a fifth aspect, the present application further provides a computer program product, which includes a computer program that implements the steps of any of the above method embodiments when executed by a processor.

[0019] The above-mentioned belt offset detection method, device, equipment and storage medium determine the target ellipse point set corresponding to the roller in the reference scale belt conveyor image, determine the roller edge midpoint based on the target ellipse point set; determine the belt edge straight line based on the reference scale belt conveyor image, calculate the projection point set of the belt edge straight line on the target ellipse point set; determine the belt edge midpoint based on the projection point set; determine the offset detection result based on the roller edge midpoint and the belt edge midpoint. Through the above implementation, first determine the target ellipse point set corresponding to the roller edge and the roller edge midpoint, then determine the projection area of the belt edge straight line on the target ellipse point set, that is, the projection point set, so that the belt edge midpoint corresponding to the roller can be determined; then, by comparing the roller edge midpoint with the belt edge midpoint, it can be determined whether the roller is offset; this application can realize belt offset detection only by collecting images and calculating and comparing the roller edge midpoint and the belt edge midpoint in the image, thereby improving the convenience of belt offset detection.

[0020] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 This is a flow chart of a belt deviation detection method provided in an embodiment of the present application;

[0023] Figure 2 A reference scale belt conveyor image provided in an embodiment of the present application;

[0024] Figure 3 This is a structural schematic diagram of a belt deviation detection device provided in an embodiment of the present application;

[0025] Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present application;

[0026] Figure 5 This is a diagram of the internal structure of a computer-readable storage medium provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the present disclosure more clearly understood, the present disclosure is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present disclosure and are not intended to limit the present disclosure.

[0028] It should be noted that the terms "first," "second," and the like in the specification and claims herein and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0029] In this document, the term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" could mean: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document generally indicates an "or" relationship between the related objects.

[0030] Example 1

[0031] Figure 1 This is a flow chart of a belt deviation detection method provided in Example 1 of the present application, refer to Figure 1 The method may be performed by a device for performing the method, and the device may be implemented by software and / or hardware. The method includes:

[0032] S110 , determining a target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determining the midpoint of the roller edge based on the target ellipse point set.

[0033] Among them, in the scenario of material transportation, a belt conveyor is often used. The belt conveyor includes a body, a row of rollers are rotatably connected to the body, and a belt is provided on the rollers; the rollers drive the belt to move during the rotation process; due to some unfavorable factors, the belt may deviate, so it is necessary to detect whether the belt is deviated to prevent the adverse effects caused by the deviation; this embodiment detects whether the belt is deviated by image processing, and for this purpose a camera is provided on the side of the belt conveyor, which is used to collect information such as Figure 2 The original image of the belt conveyor shown in the figure contains not only the belt conveyor content, but also the background content (the gray part on the right side of the belt conveyor content). However, the background content is useless for belt deviation detection, so it is necessary to delete the background content first.

[0034] Specifically, this embodiment is provided with a region of interest extraction network, which is used to extract the belt conveyor content (region of interest, ROI) in the original belt conveyor image; and the extracted belt conveyor content is recorded as the reference scale belt conveyor image.

[0035] The image of the belt conveyor with the standard scale contains rollers, such as Figure 2 As shown by the red cylinder in Figure 2 From the perspective shown, the circular side edge of the roller is elliptical; by performing image processing on the reference scale belt conveyor image, a set of edge point sets distributed along the ellipse can be obtained, and the edge point set is recorded as the target ellipse point set.

[0036] Among them, by performing corresponding mathematical calculations on the target ellipse point set, the target ellipse point set along the edge of the ellipse close to the belt side ( Figure 2 The midpoint of each point of the blue arc) distribution is recorded as the midpoint of the roller edge, such as Figure 2 As shown by the blue dots on the middle blue arc.

[0037] S120: Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set.

[0038] Among them, by performing image processing on the reference scale belt conveyor image, the belt edge points can also be determined. Each belt edge point can form a straight line, that is, the belt edge straight line, such as Figure 2 As shown by the green straight line in .

[0039] It should be noted that, in addition to determining the midpoint of the roller edge, this embodiment also seeks to find a special belt edge point on the belt edge straight line. This special belt edge point is used to compare with the above-mentioned roller edge midpoint to determine whether the belt is offset. The method of determining the special belt edge point in this embodiment is: first projecting the straight line onto the target ellipse point set to determine the corresponding projection area, which is also the projection area. Figure 2 The orange line segment is on the green line, and the set of belt edge points located in the orange line segment is recorded as the projection point set.

[0040] S130. Determine the midpoint of the belt edge based on the projection point set.

[0041] Among them, the corresponding mathematical calculation is performed on the projection point set to determine the special belt edge point in the projection point set, and the special belt edge point is recorded as the belt edge midpoint, such as Figure 2 As shown by the orange dots in the

[0042] S140. Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0043] The point distance between the midpoint of the roller edge and the midpoint of the belt edge can be calculated first, and the offset detection result can be further determined based on the point distance. The offset detection result is used to indicate whether the belt is offset and the direction of the belt offset.

[0044] Specifically, the point distance between the midpoint of the edge of the roller and the midpoint of the edge of the belt is calculated; in response to the point distance being within a preset point distance threshold range, the offset detection result is determined to be that the belt has not offset; wherein the point distance threshold range is set based on historical experience data; illustratively, the point distance threshold range is [2.3cm, 2.7cm]; in response to the point distance being greater than the point distance threshold range, the direction of the belt offset is determined to be to the left; in response to the point distance being less than the point distance threshold range, the direction of the belt offset is determined to be to the right.

[0045] It should be noted that, in this embodiment, images are only collected on one side of the belt conveyor. In another embodiment, images are collected on both sides of the belt conveyor. According to the above steps, the point distances corresponding to each image can be calculated, which are respectively recorded as the first point distance and the second point distance. If the first point distance is not equal to the second point distance, it is determined that the belt is offset, and the direction of the belt offset is toward the side of the belt conveyor corresponding to the smaller point distance between the first point distance and the second point distance.

[0046] It should be noted that this embodiment determines the target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determines the roller edge midpoint based on the target ellipse point set; determines the belt edge straight line based on the reference scale belt conveyor image, and calculates the projection point set of the belt edge straight line on the target ellipse point set; determines the belt edge midpoint based on the projection point set; and determines the offset detection result based on the roller edge midpoint and the belt edge midpoint. Through the above implementation, the target ellipse point set corresponding to the roller edge and the roller edge midpoint are first determined, and then the projection area of the belt edge straight line on the target ellipse point set, that is, the projection point set, is determined, so that the belt edge midpoint corresponding to the roller can be determined; then, by comparing the roller edge midpoint with the belt edge midpoint, it can be determined whether the roller is offset; this application can realize belt offset detection only by collecting images and calculating and comparing the roller edge midpoint and the belt edge midpoint in the image, thereby improving the convenience of belt offset detection.

[0047] Example 2

[0048] A belt deviation detection method provided in a second embodiment of the present application optimizes the step of "determining a target ellipse point set corresponding to the idler in a reference scale belt conveyor image" in the first embodiment. It should be noted that for portions not described in detail in this embodiment, reference may be made to the descriptions of other embodiments. The method includes:

[0049] S211 . Generate a multi-scale image set based on the reference-scale conveyor belt image, where the multi-scale image set includes conveyor belt images at at least two resolution scales.

[0050] Among them, the reference-scale belt conveyor image is a belt conveyor image at a normal resolution scale taken by the camera, and the normal resolution is a customized resolution of the camera; a set of ellipse point sets determined by the reference-scale belt conveyor image is subsequently used to calculate the midpoint of the roller edge, but the points in the ellipse point set determined by the reference-scale belt conveyor image may not be well distributed along the corresponding ellipse. For this reason, belt conveyor images at other resolution scales can be generated from the reference-scale belt conveyor image, and then the ellipse point sets corresponding to the belt conveyor images at multiple other resolution scales are calculated one by one, so that a set of ellipse point sets with the best distribution effect along the corresponding ellipse can be determined from the multiple ellipse point sets as the target ellipse point set.

[0051] In order to generate belt conveyor images at other resolution scales, this embodiment specifically adopts a bilinear interpolation method for processing the base-scale belt conveyor image, thereby generating multiple belt conveyor images at other resolution scales, and recording the multiple belt conveyor images at other resolution scales as a multi-scale image set; illustratively, the base-scale belt conveyor image is bilinearly interpolated down-sampled to 0.5 times the resolution to obtain a 0.5 times resolution belt conveyor image, and the base-scale belt conveyor image is bilinearly interpolated up-sampled to 2 times the resolution to obtain a 2 times resolution belt conveyor image; the base-scale belt conveyor image (1 times resolution belt conveyor image), the 0.5 times resolution belt conveyor image, and the 2 times resolution belt conveyor image are combined into a multi-scale image set.

[0052] S212: Determine initial ellipse point sets corresponding to the respective conveyor belt images in the multi-scale image set.

[0053] Among them, this embodiment presets an ellipse fitting algorithm, which is used to process each belt conveyor image in the multi-scale image set, and output an ellipse point set corresponding to each belt conveyor image one by one, and the ellipse point set output by the ellipse fitting algorithm is recorded as the initial ellipse point set; illustratively, the ellipse fitting algorithm used in this embodiment is the Fitzgibbon algorithm, and in other embodiments, there is no specific limitation.

[0054] S213 , calculating the gradient consistency index corresponding to each of the initial ellipse point sets.

[0055] Among them, taking one of the initial ellipse point sets as an example, the corresponding gradient consistency index S can be obtained by further mathematical calculation of the initial ellipse point set. The gradient consistency index S is used to reflect the distribution effect of each point in the corresponding initial ellipse point set along the corresponding ellipse.

[0056] S214 : Determine a candidate ellipse point set based on the gradient consistency index and the index threshold.

[0057] Among them, the indicator threshold is set according to historical experience data. In this embodiment, the indicator threshold is 0.8. In other embodiments, it is not specifically limited. The indicator threshold is used for comparison with the gradient consistency index S. If the gradient consistency index S is greater than the indicator threshold, it means that the distribution of each point along the corresponding ellipse in the initial ellipse point set corresponding to the gradient consistency index S is better, and the initial ellipse also has the potential to become the target ellipse point set; and the initial ellipse point set whose gradient consistency index S is greater than the indicator threshold is recorded as a candidate ellipse point set.

[0058] S215 : In response to the number of the candidate ellipse point sets being greater than one, calculating the point set scores corresponding to the candidate ellipse point sets.

[0059] Among them, if the number of candidate ellipse point sets is one, the candidate ellipse point set can be directly used as the target ellipse point set; if the number of candidate ellipse point sets is greater than one, then the target ellipse point set needs to be selected from each candidate ellipse point set; in this embodiment, the candidate ellipse point set is further considered from multiple dimensions, so as to calculate the total score of the candidate ellipse point set in multiple dimensions, and the total score is recorded as the point set score Score.

[0060] S216 : Determine a target ellipse point set based on the candidate ellipse point sets and the point set scores.

[0061] Among them, the target ellipse point set is the candidate ellipse point set corresponding to the maximum point set score.

[0062] S217: Determine the midpoint of the roller edge based on the target ellipse point set.

[0063] S220: Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set.

[0064] S230: Determine the midpoint of the belt edge based on the projection point set.

[0065] S240: Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0066] Example 3

[0067] A belt deviation detection method provided in a third embodiment of the present application optimizes the step of "calculating the gradient consistency index corresponding to each of the initial ellipse point sets" in the second embodiment. It should be noted that for portions not described in detail in this embodiment, reference may be made to the descriptions of other embodiments. The method includes:

[0068] S311 . Generate a multi-scale image set based on the reference-scale conveyor belt image, where the multi-scale image set includes conveyor belt images at at least two resolution scales.

[0069] S312: Determine initial ellipse point sets corresponding to the respective conveyor belt images in the multi-scale image set.

[0070] S313A: Convert each of the initial ellipse point sets to a corresponding target coordinate system to obtain a converted ellipse point set corresponding to each of the initial ellipse point sets.

[0071] Among them, taking one of the initial ellipse point sets as an example, the ellipse corresponding to the initial ellipse point set can be calculated through the ellipse fitting algorithm. The ellipse has its major axis and minor axis. A new coordinate system is established with the major axis as the X-axis and the minor axis as the Y-axis, and the new coordinate system is recorded as the target coordinate system; through the corresponding coordinate system conversion matrix, the initial ellipse point set can be converted from the original coordinate system to the target coordinate system, and the new point set obtained by converting the initial ellipse point set to the target coordinate system is recorded as the converted ellipse point set.

[0072] S313B, calculate the local curvature corresponding to each transformed ellipse point set.

[0073] Among them, taking one of the converted ellipse point sets as an example, the converted ellipse point set corresponds to an ellipse in the target coordinate system, and each point in the converted ellipse point set has a corresponding point in the ellipse, and the set of corresponding points is recorded as the ideal ellipse point set; it should be noted that each point in the converted ellipse point set may be located on the corresponding ellipse or may not be located on the corresponding ellipse; each point in the ideal ellipse point set is located on the corresponding ellipse; the ellipse curvature radius can be obtained by performing corresponding calculations on the converted ellipse point set and the corresponding ideal ellipse point set, and the local curvature is the inverse of the ellipse curvature radius.

[0074] S313C: In response to the local curvature not exceeding the preset curvature threshold, performing break compensation on the converted ellipse point sets corresponding to the local curvatures, to obtain compensated ellipse point sets corresponding to the converted ellipse point sets.

[0075] Here, the local curvature corresponding to one of the converted ellipse point sets is used as an example. The local curvature is used to indicate how well the corresponding converted ellipse point set fits the ellipse (is distributed along the ellipse). To facilitate measuring this local curvature, this embodiment sets a curvature threshold based on historical empirical data. If the local curvature does not exceed this curvature threshold, it indicates that the corresponding converted ellipse point set has a good ellipse fitting effect.

[0076] It should be noted that the gap between two adjacent points in the converted ellipse point set may be too large. This large gap can also be regarded as a break in the ellipse corresponding to the converted ellipse point set. For this reason, it is necessary to perform break compensation on the converted ellipse point set. Break compensation is to perform lateral compensation on the two points (break points) involved in the break, and the new point set obtained after the break compensation of the converted ellipse point set is recorded as the compensated ellipse point set.

[0077] S313D. Calculate the gradient consistency index corresponding to each compensated ellipse point set.

[0078] Among them, taking one of the compensated ellipse point sets as an example, the compensated ellipse point set contains N points (x i ,yi ), each point is processed separately by the preset Sobel operator, and the gradient direction angle corresponding to each point can be obtained. Each point has its corresponding ellipse geometric normal direction angle φ normal (x i ,y i ); The calculation formula of the gradient consistency index S of the compensated ellipse point set is as follows:

[0079]

[0080] S314 : Determine a candidate ellipse point set based on the gradient consistency index and the index threshold.

[0081] S315 : In response to the number of the candidate ellipse point sets being greater than one, calculating the point set scores corresponding to the candidate ellipse point sets.

[0082] S316 : Determine a target ellipse point set based on the candidate ellipse point sets and the point set scores.

[0083] S317: Determine the midpoint of the roller edge based on the target ellipse point set.

[0084] S320: Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set.

[0085] S330: Determine the midpoint of the belt edge based on the projection point set.

[0086] S340. Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0087] Example 4

[0088] A belt deviation detection method is provided in a fourth embodiment of the present application. This method optimizes the method of "performing breakage compensation on each converted ellipse point set corresponding to each local curvature to obtain a compensated ellipse point set corresponding to each converted ellipse point set" in the third embodiment. It should be noted that for portions not described in detail in this embodiment, reference may be made to the descriptions of other embodiments. This method includes:

[0089] S411 . Generate a multi-scale image set based on the reference-scale conveyor belt image, where the multi-scale image set includes conveyor belt images at at least two resolution scales.

[0090] S412: Determine initial ellipse point sets corresponding to the conveyor belt images in the multi-scale image set.

[0091] S413A: Convert each of the initial ellipse point sets to a corresponding target coordinate system to obtain a converted ellipse point set corresponding to each of the initial ellipse point sets.

[0092] S413B, calculate the local curvature corresponding to each transformed ellipse point set.

[0093] S413C1. In response to the local curvature not exceeding the preset curvature threshold, the perimeter continuity of each of the converted ellipse point sets is calculated; in response to the perimeter continuity being less than the continuity threshold, the fracture offset corresponding to the converted ellipse point set is calculated; or, the cumulative fracture length of each of the converted ellipse point sets is calculated; in response to the cumulative fracture length being greater than the length threshold, the fracture offset corresponding to the converted ellipse point set is calculated.

[0094] Among them, if the local curvature does not exceed the preset curvature threshold, it means that the ellipse fitting effect of the converted ellipse point set corresponding to the local curvature meets the standard; at this time, it is necessary to further determine whether the converted ellipse point set needs to be compensated for fractures, because the converted ellipse point set may have serious fractures; for example, if the sum of the distances between two adjacent points with large spacing in the converted ellipse point set is too large, or there are many points in the converted ellipse point set that do not fall on the ellipse corresponding to the converted ellipse point set, then it is determined that the converted ellipse point set has serious fractures.

[0095] This embodiment provides two methods for determining whether the ellipse point set is severely broken after conversion:

[0096] First, the sum of the distances between two adjacent points with the largest spacing in the ellipse point set after the conversion is calculated to obtain the cumulative length L of the fracture. break , where if the distance between two adjacent points exceeds the custom threshold, it is considered that the distance between the two adjacent points is large; and the sum of the distances between each adjacent point in the converted ellipse point set is calculated as L total ; Then calculate the perimeter continuity C corresponding to the transformed ellipse point set cont , where the perimeter continuity C cont The calculation formula is:

[0097] It should be noted that this embodiment is aimed at the perimeter continuity C cont A continuity threshold is set, if the perimeter continuity C cont If the continuity threshold is less than the threshold, it is determined that the converted ellipse point set corresponding to the perimeter continuity is severely broken.

[0098] Secondly, the sum of the distances between two adjacent points with the largest spacing in the ellipse point set after the conversion is calculated to obtain the cumulative length L of the fracture. break It should be noted that this embodiment is directed to the cumulative length L of the fracture. breakA corresponding length threshold is set. If the cumulative length of the fracture is L break If the length is greater than the threshold, it is determined that the converted ellipse point set corresponding to the perimeter continuity is severely broken.

[0099] It should be noted that any of the above two methods can be used to determine whether the converted ellipse point set is severely broken; if it is determined that the converted ellipse point set is severely broken, the converted ellipse point set needs to be compensated for the break; the corresponding compensation amount can be calculated based on the converted ellipse point set, and the compensation amount is recorded as the break offset.

[0100] S413C2. Compensate the corresponding converted ellipse point sets based on the respective fracture offsets to obtain compensated ellipse point sets corresponding to the respective converted ellipse point sets.

[0101] Among them, taking the fracture offset corresponding to one of the converted ellipse point sets as an example, the fracture offset is used to perform lateral compensation on the fracture points in the converted ellipse point set, and the new point set obtained after the compensation of the converted ellipse point set is recorded as the compensated ellipse point set.

[0102] S413D. Calculate the gradient consistency index corresponding to each compensated ellipse point set.

[0103] S414 : Determine a candidate ellipse point set based on the gradient consistency index and the index threshold.

[0104] S415 : In response to the number of the candidate ellipse point sets being greater than one, calculating the point set scores corresponding to the candidate ellipse point sets.

[0105] S416 : Determine a target ellipse point set based on the candidate ellipse point sets and the point set scores.

[0106] S417: Determine the midpoint of the roller edge based on the target ellipse point set.

[0107] S420: Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set.

[0108] S430: Determine the midpoint of the belt edge based on the projection point set.

[0109] S440: Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0110] Example 5

[0111] A belt deviation detection method provided in a fifth embodiment of the present application optimizes the step of "calculating the point set scores corresponding to each candidate ellipse point set" in the second embodiment. It should be noted that for portions not described in detail in this embodiment, reference may be made to the descriptions of other embodiments. The method includes:

[0112] S511 . Generate a multi-scale image set based on the reference-scale conveyor belt image, where the multi-scale image set includes conveyor belt images at at least two resolution scales.

[0113] S512: Determine initial ellipse point sets corresponding to the respective conveyor belt images in the multi-scale image set.

[0114] S513 , calculating the gradient consistency index corresponding to each of the initial ellipse point sets.

[0115] S514 : Determine a candidate ellipse point set based on the gradient consistency index and the index threshold.

[0116] S515A: In response to the number of the candidate ellipse point sets being greater than one, determine the proportion of matching edge points, perimeter continuity, and the gradient consistency index corresponding to each of the candidate ellipse point sets.

[0117] Among them, the number of points in the candidate ellipse point set that fall on the ellipse corresponding to the candidate ellipse point set is counted to obtain the number of matching edge points N match ; Then, determine the total number of points N in the candidate ellipse point set total , and finally calculate the proportion of matching edge points C cov , where the number of matching edge points accounts for C cov The calculation formula is: The sum of the distances between two adjacent points with the largest spacing in the candidate ellipse point set is also counted to obtain the cumulative length L of the fracture. break , where the distance between two adjacent points exceeds the custom threshold, it is considered that the distance between the two adjacent points is large; and the sum of the distances between each two adjacent points in the candidate ellipse point set is calculated L total ; Then calculate the perimeter continuity C corresponding to the candidate ellipse point set cont , where the perimeter continuity C cont The calculation formula is: Each candidate ellipse point set has a corresponding gradient consistency index S. The calculation method of the gradient consistency index S is shown in the above embodiment and will not be described in detail.

[0118] S515B. Based on the proportion of matching edge points, the perimeter continuity and the gradient consistency index corresponding to each of the candidate ellipse point sets, and the weight parameter group, obtain the point set score corresponding to each of the candidate ellipse point sets; wherein the weight parameter group includes weight parameters corresponding to the proportion of matching edge points, the perimeter continuity and the gradient consistency index, respectively.

[0119] Among them, the weight parameter group is the proportion of matching edge points C cov The weight parameter set is τ, the weight parameter set for the gradient consistency index S is β, and the weight parameter set for the perimeter continuity C is cont The weight parameter is set to γ; in this embodiment, τ=0.4, β=0.3, and γ=0.3; in other embodiments, the specific parameters are not limited.

[0120] The calculation formula of the point set score is as follows:

[0121] Score=τ·C cov +β·S+γ·C cont .

[0122] S516 : Determine a target ellipse point set based on the candidate ellipse point sets and the point set scores.

[0123] S517: Determine the midpoint of the roller edge based on the target ellipse point set.

[0124] S520: Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set.

[0125] S530: Determine the midpoint of the belt edge based on the projection point set.

[0126] S540: Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0127] Example 6

[0128] A belt deviation detection method is provided in a sixth embodiment of the present application. This method optimizes the "determining the midpoint of the roller edge based on the target ellipse point set" in the first embodiment. It should be noted that for portions not described in detail in this embodiment, reference may be made to the descriptions of other embodiments. This method includes:

[0129] S611. Determine a target ellipse point set corresponding to the roller in the reference scale belt conveyor image.

[0130] S612: Convert the target ellipse point set into a polar coordinate system to obtain a target polar coordinate point set.

[0131] Among them, the new point set obtained by converting the target ellipse point set into the polar coordinate system is recorded as the target polar coordinate point set; each polar coordinate point in the target polar coordinate point set corresponds one to one to each rectangular coordinate system point in the target ellipse point set.

[0132] S613: Generate an angle distribution histogram based on the target polar coordinate point set.

[0133] Among them, each polar coordinate in the target polar coordinate point set has its corresponding polar angle. In this embodiment, multiple polar angle intervals are set for each polar angle. By placing each polar angle corresponding to the target polar coordinate point set in the corresponding polar angle interval, the number of points in each polar angle interval is counted, and an angle distribution histogram can be generated.

[0134] S614: Determine a maximum continuous interval based on the angle distribution histogram.

[0135] The maximum continuous interval is the longest continuous interval in the angle distribution histogram with a number of points greater than 0.

[0136] S615: Determine a central angle based on the maximum continuous interval.

[0137] Among them, the central angle α mid It is half of the sum of the maximum polar angle and the minimum polar angle in the maximum continuous interval.

[0138] S616. Determine the midpoint of the edge of the roller based on the central angle and the gradient direction compensation coefficient, gradient direction angle, and unit normal direction vector corresponding to the central angle.

[0139] Among them, the central angle α mid There is a corresponding gradient direction compensation coefficient λ. In this embodiment, the value range of the gradient direction compensation coefficient λ is 0.1-0.3. The central angle α is calculated by Sobel operator. mid The central angle α can be calculated by calculation mid The corresponding gradient direction angle The central angle α mid The corresponding unit normal direction vector is recorded as The roller edge midpoint M can be calculated through the center angle and the gradient direction compensation coefficient corresponding to the center angle, the gradient direction angle, and the unit normal direction vector. roller , where the midpoint M of the roller edge roller The calculation formula is as follows:

[0140]

[0141] Among them, φ(α mid ) is the central angle α mid The corresponding rectangular coordinates.

[0142] S620: Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set.

[0143] S630: Determine the midpoint of the belt edge based on the projection point set.

[0144] S640: Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0145] Example 7

[0146] A belt deviation detection method provided in Example 7 of the present application optimizes the "determining the belt edge straight line based on the reference scale belt conveyor image" in Example 1 or 2. It should be noted that for parts not described in detail in this embodiment, reference can be made to the descriptions of other embodiments. The method includes:

[0147] S710: Determine a target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determine the midpoint of the roller edge based on the target ellipse point set.

[0148] S721 , calculating edge segment sets corresponding to the belt conveyor images at different resolution scales in the multi-scale image set corresponding to the reference-scale belt conveyor image.

[0149] Among them, taking the belt conveyor image at a resolution scale in the multi-scale image set as an example, this embodiment presets a line segment detection algorithm. Exemplarily, the line segment detection algorithm is the LSD algorithm. The line segment detection algorithm processes the belt conveyor image at a resolution scale and can output multiple line segments located at the edge position of the belt in the belt conveyor image. The multiple line segments are distributed along the straight line where the belt edge is located; and the combination of the multiple line segments is recorded as an edge line segment set.

[0150] S722: Map the edge line segment sets corresponding to the belt conveyor images at other resolution scales to the edge line segment set corresponding to the belt conveyor image at the reference scale to obtain a comprehensive line segment set.

[0151] Among them, the line segment detection algorithm processes the belt conveyor image at the base scale to obtain the edge line segment set corresponding to the belt conveyor image at the base scale, and the edge line segment set corresponding to the belt conveyor image at the base scale is recorded as the base edge line segment set; the line segment detection algorithm also processes the belt conveyor images at other resolution scales one by one, thereby obtaining edge line segment sets that correspond one to one with the belt conveyor images at other resolution scales; and the edge line segment sets that correspond one to one with the belt conveyor images at other resolution scales are all mapped to the base edge line segment set, thereby obtaining a new line segment set, and the new line segment set is recorded as the comprehensive line segment set.

[0152] S723: Calculate the line segment weights corresponding to the respective line segments in the comprehensive line segment set, and determine the overlapping line segment groups in the comprehensive line segment set.

[0153] Among them, the calculation formula of the line segment weight W is: W=M 2 / (d+ε), where M is the gradient strength, d is the line segment density, and ε is a constant. The Sobel operator processes the line segment to obtain the gradient corresponding to each pixel in the line segment. The gradient strength M is calculated by summing the squares of the gradients and then taking the square root of the square sum. Taking a line segment in the comprehensive line segment set as an example, the number of line segments in an area near the line segment is the line segment density d corresponding to the line segment.

[0154] It should be noted that the comprehensive line segment set contains the line segments corresponding to the belt conveyor images at various resolution scales. Therefore, for a position on the belt edge, multiple overlapping line segments may appear, and multiple overlapping line segments correspond to belt conveyor images at multiple different resolution scales; and the set of line segments in the comprehensive line segment set that correspond to the same position on the belt edge and overlap is recorded as an overlapping line segment group.

[0155] S724: Retain the line segment with the largest line segment weight in each overlapping line segment group to obtain the belt edge line. Each line segment in the overlapping line segment group has a corresponding line segment weight. The line segment with the largest line segment weight in each overlapping line segment group is retained, and then the remaining line segments in the overlapping line segment group that do not have the largest line segment weight are deleted. The remaining line segments in the comprehensive line segment set are recorded as the belt edge line.

[0156] S725: Calculate the projection point set of the belt edge straight line on the target ellipse point set.

[0157] S730: Determine the midpoint of the belt edge based on the projection point set.

[0158] S740: Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0159] Example 8

[0160] A belt deviation detection method is provided in an eighth embodiment of the present application. This method optimizes the "determining the belt edge midpoint based on the projection point set" in the first embodiment. It should be noted that for portions not described in detail in this embodiment, reference may be made to the descriptions in other embodiments. The method includes:

[0161] S810: Determine a target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determine the midpoint of the roller edge based on the target ellipse point set.

[0162] S820: Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set.

[0163] S831. Convert the projected point set to the polar coordinate system where the target ellipse point set is located to obtain a projected polar coordinate point set.

[0164] Among them, the new point set obtained by converting the projection point set into the polar coordinate system is also the projection polar coordinate point set; the polar coordinates in the projection polar coordinate point set and the rectangular coordinates (x i ,y i ) correspond one to one.

[0165] S832: Calculate the Gaussian neighborhood edge weight corresponding to each point in the projected polar coordinate point set.

[0166] Among them, each point in the projected polar coordinate point set (polar coordinates, (r i ,θ i )) can calculate the corresponding Gaussian neighborhood edge weight W(x i ,y i ), where the Gaussian neighborhood edge weight W(x i ,y i ) is calculated as follows:

[0167]

[0168] Among them, M(x i ,y i ) is the gradient intensity corresponding to the point (polar coordinate) in the projected polar coordinate point set, r i is the pole radius of the point, a is the minor axis length of the ellipse corresponding to the target ellipse point set, and σ is the Gaussian bandwidth parameter (constant).

[0169] S833. Determine the midpoint of the belt edge based on the Gaussian neighborhood edge weights and the projected polar coordinate point set.

[0170] Among them, the midpoint M of the belt edge belt The calculation formula is:

[0171]

[0172] Among them, x i is the rectangular coordinate abscissa corresponding to the polar coordinates of the projected polar coordinate point set, y i It is the rectangular coordinate ordinate corresponding to the polar coordinate in the projected polar coordinate point set.

[0173] S840. Determine an offset detection result based on the midpoint of the roller edge and the midpoint of the belt edge.

[0174] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0175] Embodiment 9

[0176] Based on the same inventive concept, this embodiment also provides a belt deviation detection device for implementing the aforementioned belt deviation detection method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more belt deviation detection device embodiments provided below can be found in the aforementioned limitations of the belt deviation detection method and will not be further elaborated here.

[0177] In this embodiment, Figure 3 As shown, a belt deviation detection device is provided, comprising:

[0178] A roller midpoint determination module is used to determine a target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determine the roller edge midpoint based on the target ellipse point set;

[0179] A projection point set generating module is used to determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set;

[0180] a belt midpoint determination module, configured to determine a belt edge midpoint based on the projection point set;

[0181] The detection result generating module is used to determine the offset detection result based on the midpoint of the edge of the roller and the midpoint of the edge of the belt.

[0182] Each module in the belt deviation detection device described above can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0183] It should be noted that this embodiment determines the target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determines the roller edge midpoint based on the target ellipse point set; determines the belt edge straight line based on the reference scale belt conveyor image, and calculates the projection point set of the belt edge straight line on the target ellipse point set; determines the belt edge midpoint based on the projection point set; and determines the offset detection result based on the roller edge midpoint and the belt edge midpoint. Through the above implementation, the target ellipse point set corresponding to the roller edge and the roller edge midpoint are first determined, and then the projection area of the belt edge straight line on the target ellipse point set, that is, the projection point set, is determined, so that the belt edge midpoint corresponding to the roller can be determined; then, by comparing the roller edge midpoint with the belt edge midpoint, it can be determined whether the roller is offset; this application can realize belt offset detection only by collecting images and calculating and comparing the roller edge midpoint and the belt edge midpoint in the image, thereby improving the convenience of belt offset detection.

[0184] In an optional embodiment, in determining the target ellipse point set corresponding to the roller in the reference scale belt conveyor image, the roller midpoint determination module is specifically configured to:

[0185] generating a multi-scale image set based on the reference-scale belt conveyor image, wherein the multi-scale image set includes belt conveyor images at at least two resolution scales;

[0186] Determine an initial ellipse point set corresponding to each of the belt conveyor images in the multi-scale image set;

[0187] Calculating the gradient consistency index corresponding to each of the initial ellipse point sets;

[0188] Determining a candidate ellipse point set based on the gradient consistency index and the index threshold;

[0189] In response to the number of the candidate ellipse point sets being greater than one, calculating a point set score corresponding to each of the candidate ellipse point sets;

[0190] Based on the candidate ellipse point sets and the scores of the point sets, a target ellipse point set is determined.

[0191] In an optional embodiment, in terms of calculating the gradient consistency index corresponding to each of the initial ellipse point sets, the roller midpoint determination module is specifically configured to:

[0192] Converting each of the initial ellipse point sets to a corresponding target coordinate system to obtain converted ellipse point sets corresponding to each of the initial ellipse point sets;

[0193] Calculate the local curvature corresponding to each transformed ellipse point set;

[0194] In response to the local curvature not exceeding a preset curvature threshold, performing break compensation on the converted ellipse point sets corresponding to the local curvatures, respectively, to obtain compensated ellipse point sets corresponding to the converted ellipse point sets;

[0195] Calculate the gradient consistency index corresponding to each compensated ellipse point set.

[0196] In an optional embodiment, in terms of performing fracture compensation on the converted ellipse point sets corresponding to the local curvatures to obtain compensated ellipse point sets corresponding to the converted ellipse point sets, the roller midpoint determination module is specifically configured to:

[0197] Calculating the perimeter continuity of each of the converted ellipse point sets, and in response to the perimeter continuity being less than a continuity threshold, calculating the break offset corresponding to the converted ellipse point set; or calculating the cumulative break length of each of the converted ellipse point sets, and in response to the cumulative break length being greater than a length threshold, calculating the break offset corresponding to the converted ellipse point set;

[0198] The corresponding converted ellipse point sets are compensated based on the respective fracture offsets to obtain compensated ellipse point sets corresponding to the respective converted ellipse point sets.

[0199] In an optional embodiment, in terms of calculating the point set scores corresponding to each of the candidate ellipse point sets, the roller midpoint determination module is specifically configured to:

[0200] Determine the proportion of matching edge points, perimeter continuity, and gradient consistency index corresponding to each of the candidate ellipse point sets;

[0201] Based on the proportion of matching edge points, the perimeter continuity and the gradient consistency index corresponding to each of the candidate ellipse point sets, and a weight parameter group, the point set score corresponding to each of the candidate ellipse point sets is obtained; wherein the weight parameter group includes weight parameters corresponding to the proportion of matching edge points, the perimeter continuity and the gradient consistency index, respectively.

[0202] In an optional embodiment, in determining the midpoint of the roller edge based on the target ellipse point set, the roller midpoint determination module is specifically configured to:

[0203] Converting the target ellipse point set to a polar coordinate system to obtain a target polar coordinate point set;

[0204] generating an angle distribution histogram based on the target polar coordinate point set;

[0205] determining a maximum continuous interval based on the angle distribution histogram;

[0206] determining a central angle based on the maximum continuous interval;

[0207] The midpoint of the edge of the roller is determined based on the central angle and the gradient direction compensation coefficient, the gradient direction angle, and the unit normal direction vector corresponding to the central angle.

[0208] In an optional embodiment, in determining the belt edge straight line based on the reference scale belt conveyor image, the projection point set generation module is specifically configured to:

[0209] Calculate the edge segment sets corresponding to the belt conveyor images at each resolution scale in the multi-scale image set corresponding to the reference-scale belt conveyor image;

[0210] Mapping the edge line segment sets corresponding to the belt conveyor images at other resolution scales to the edge line segment set corresponding to the belt conveyor image at the reference scale to obtain a comprehensive line segment set;

[0211] Calculating the line segment weights corresponding to the respective line segments in the comprehensive line segment set, and determining overlapping line segment groups in the comprehensive line segment set;

[0212] The line segment with the largest line segment weight in each of the overlapping line segment groups is retained to obtain the belt edge straight line.

[0213] In an optional embodiment, in determining the midpoint of the belt edge based on the projection point set, the belt midpoint determination module is specifically configured to:

[0214] Converting the projected point set to the polar coordinate system where the target ellipse point set is located to obtain a projected polar coordinate point set;

[0215] Calculate the Gaussian neighborhood edge weight corresponding to each point in the projected polar coordinate point set;

[0216] The midpoint of the belt edge is determined based on the edge weights of each of the Gaussian neighborhoods and the projected polar coordinate point set.

[0217] Example 10

[0218] In this embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a belt deviation detection method is implemented.

[0219] Those skilled in the art will understand that Figure 4 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present disclosure, and does not constitute a limitation on the computer device to which the solution of the present disclosure is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0220] Example 11

[0221] In this embodiment, a computer readable storage medium is provided. Figure 5 As shown, a computer program is stored thereon, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0222] Example 12

[0223] In this embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0224] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0225] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in the present disclosure may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this disclosure may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in each embodiment provided in this disclosure may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, etc.

[0226] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0227] The above-described embodiments merely represent several implementation methods of the present disclosure. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present disclosure. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present disclosure, all of which fall within the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure shall be determined by the appended claims.

Claims

1. A belt deviation detection method, characterized in that: include: Determine a target ellipse point set corresponding to the idler in the reference scale belt conveyor image, and determine the midpoint of the idler edge based on the target ellipse point set; Determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set; determining a belt edge midpoint based on the projected point set; An offset detection result is determined based on the midpoint of the roller edge and the midpoint of the belt edge.

2. The method according to claim 1, characterized in that The step of determining a target ellipse point set corresponding to the roller in the reference scale belt conveyor image includes: generating a multi-scale image set based on the reference-scale belt conveyor image, wherein the multi-scale image set includes belt conveyor images at at least two resolution scales; Determine an initial ellipse point set corresponding to each of the belt conveyor images in the multi-scale image set; Calculating the gradient consistency index corresponding to each of the initial ellipse point sets; Determining a candidate ellipse point set based on the gradient consistency index and the index threshold; In response to the number of the candidate ellipse point sets being greater than one, calculating a point set score corresponding to each of the candidate ellipse point sets; Based on the candidate ellipse point sets and the scores of the point sets, a target ellipse point set is determined.

3. The method according to claim 2, characterized in that The calculating of the gradient consistency index corresponding to each of the initial ellipse point sets includes: Converting each of the initial ellipse point sets to a corresponding target coordinate system to obtain converted ellipse point sets corresponding to each of the initial ellipse point sets; Calculate the local curvature corresponding to each transformed ellipse point set; In response to the local curvature not exceeding a preset curvature threshold, performing break compensation on the converted ellipse point sets corresponding to the local curvatures, respectively, to obtain compensated ellipse point sets corresponding to the converted ellipse point sets; Calculate the gradient consistency index corresponding to each compensated ellipse point set.

4. The method according to claim 3, characterized in that The performing break compensation on the converted ellipse point sets corresponding to the local curvatures one by one to obtain compensated ellipse point sets corresponding to the converted ellipse point sets respectively includes: Calculating the perimeter continuity of each of the converted ellipse point sets, and in response to the perimeter continuity being less than a continuity threshold, calculating the break offset corresponding to the converted ellipse point set; or calculating the cumulative break length of each of the converted ellipse point sets, and in response to the cumulative break length being greater than a length threshold, calculating the break offset corresponding to the converted ellipse point set; The corresponding converted ellipse point sets are compensated based on the respective fracture offsets to obtain compensated ellipse point sets corresponding to the respective converted ellipse point sets.

5. The method according to claim 2, characterized in that Calculating the point set scores corresponding to each of the candidate ellipse point sets includes: Determine the proportion of matching edge points, perimeter continuity, and gradient consistency index corresponding to each of the candidate ellipse point sets; Based on the proportion of matching edge points, the perimeter continuity and the gradient consistency index corresponding to each of the candidate ellipse point sets, and a weight parameter group, the point set score corresponding to each of the candidate ellipse point sets is obtained; wherein the weight parameter group includes weight parameters corresponding to the proportion of matching edge points, the perimeter continuity and the gradient consistency index, respectively.

6. The method according to claim 1, characterized in that Determining the midpoint of the roller edge based on the target ellipse point set includes: Converting the target ellipse point set to a polar coordinate system to obtain a target polar coordinate point set; generating an angle distribution histogram based on the target polar coordinate point set; determining a maximum continuous interval based on the angle distribution histogram; determining a central angle based on the maximum continuous interval; The midpoint of the edge of the roller is determined based on the central angle and the gradient direction compensation coefficient, the gradient direction angle, and the unit normal direction vector corresponding to the central angle.

7. The method according to any one of claims 1 or 2, characterized in that The determining of the belt edge straight line based on the reference scale belt conveyor image includes: Calculate the edge segment sets corresponding to the belt conveyor images at each resolution scale in the multi-scale image set corresponding to the reference-scale belt conveyor image; Mapping the edge line segment sets corresponding to the belt conveyor images at other resolution scales to the edge line segment set corresponding to the belt conveyor image at the reference scale to obtain a comprehensive line segment set; Calculating the line segment weights corresponding to the respective line segments in the comprehensive line segment set, and determining overlapping line segment groups in the comprehensive line segment set; The line segment with the largest line segment weight in each of the overlapping line segment groups is retained to obtain the belt edge straight line.

8. The method according to claim 1, characterized in that Determining the midpoint of the belt edge based on the projection point set includes: Converting the projected point set to the polar coordinate system where the target ellipse point set is located to obtain a projected polar coordinate point set; Calculate the Gaussian neighborhood edge weight corresponding to each point in the projected polar coordinate point set; The midpoint of the belt edge is determined based on the edge weights of each of the Gaussian neighborhoods and the projected polar coordinate point set.

9. A belt deviation detection device, characterized in that: The device comprises: A roller midpoint determination module is used to determine a target ellipse point set corresponding to the roller in the reference scale belt conveyor image, and determine the roller edge midpoint based on the target ellipse point set; A projection point set generating module is used to determine a belt edge straight line based on the reference scale belt conveyor image, and calculate a projection point set of the belt edge straight line on the target ellipse point set; a belt midpoint determination module, configured to determine a belt edge midpoint based on the projection point set; The detection result generating module is used to determine the offset detection result based on the midpoint of the edge of the roller and the midpoint of the edge of the belt.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images