Jaw clutch pull-in detection method and equipment based on machine vision and medium

Abstract

The invention discloses a jaw clutch pull-in detection method and device based on machine vision and a medium, and the method comprises the following steps: obtaining an image of a tooth gap region of a jaw clutch, and carrying out the preprocessing of the image; performing edge detection on the preprocessed image by using an improved Canny operator to obtain the edge of the clutch backlash; detecting parallel straight lines in the edge image by adopting a Hough transform method, wherein the parallel straight lines are used for representing upper and lower edges of the backlash; a parallel straight line pair is screened out through a mean shift clustering method, the pixel width of the parallel straight line pair is calculated, and the pixel width is converted into the actual backlash width; and determining the pull-in state of the jaw clutch according to the actual backlash width. According to the invention, the pull-in state of the clutch can be accurately judged, and the automation and intelligence level of an industrial field can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment monitoring and control systems, and in particular to a tooth clutch engagement detection method, equipment and medium based on machine vision. Background Art

[0002] The clutch plays an important transmission role in the mechanical system to ensure the normal operation of the mechanical equipment. At present, the clutch engagement detection mainly relies on manual inspection, especially the tooth clutch is installed inside large equipment, and manual inspection has problems such as low efficiency and high risk. In addition, there is little research on the automated clutch engagement detection, which is difficult to meet the needs of industrial sites for efficient automatic detection.

[0003] Machine vision technology has been widely used in industrial inspection in recent years. It can replace manual non-contact inspection, especially in edge detection and straight line detection. This technology can effectively overcome the problems of low accuracy and poor real-time performance of manual inspection through image processing methods. Although machine vision technology has been used in many fields, its application in automatic state detection of clutches is still relatively rare. Summary of the invention

[0004] The purpose of the present invention is to provide a tooth clutch engagement detection method, device, and medium based on machine vision, which can accurately determine the clutch engagement state and help improve the automation and intelligence level of industrial sites.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for detecting engagement of a tooth clutch based on machine vision, the method comprising the following steps:

[0007] Acquire an image of the tooth gap area of ​​the dog clutch and perform preprocessing;

[0008] The edge of the clutch gear gap is obtained by using the improved Canny operator to detect the edge of the preprocessed image.

[0009] The Hough transform method is used to detect parallel straight lines in the edge image, and the parallel straight lines are used to represent the upper and lower edges of the tooth gap;

[0010] Screening out parallel straight line pairs by using a mean shift clustering method, calculating the pixel width of the parallel straight line pairs, and converting the pixel width into an actual tooth gap width;

[0011] The engagement state of the dog clutch is determined according to the actual tooth gap width.

[0012] In some embodiments, screening out parallel straight line pairs by a mean shift clustering method comprises:

[0013] Determine the weight parameters and mean shift function of the parallel straight lines, and calculate the weighted mean shift vector of each data point;

[0014] Update the data points until the weighted mean drift vector of all data points converges to the preset value and determine the cluster center;

[0015] A pair of parallel straight lines is determined according to the cluster centers.

[0016] In some embodiments, the weight parameters of the parallel straight lines are:

[0017]

[0018] Among them, i is the parallel line whose weight parameter is being calculated, n is all the parallel lines detected, and j is the parallel lines that meet the condition L i Parallel straight lines; S j is the score of the parallel line j, ρ is the perpendicular distance from the origin to the parallel line, θ is the angle of the parallel line n, S n is the score of the parallel line n, L i is an angle θ i is a cluster of straight lines under the condition, l j is the set L i An element in .

[0019] In some embodiments, the weighted mean shift vector for each data point is calculated by the following formula:

[0020]

[0021] Among them, the weight is w={w1,w2,…,w m}, In the n+1th iteration, the angular center of the kth line cluster is updated; is the current angular center of the kth line cluster in the nth iteration, is the angle parameter of the i-th line in the n-th iteration; h is the bandwidth parameter, and g(x) is the kernel function.

[0022] In some embodiments, the Hough transform method is used to detect parallel straight lines in an edge image, including:

[0023] Identify the parallel straight line edges of the tooth gap and determine the parallel characteristics of the parallel straight lines by calculating the angle and distance relationship between adjacent edge points.

[0024] In some implementations, determining the engagement state of the dog clutch according to the actual tooth gap width includes:

[0025] The actual backlash width is compared with a preset engagement threshold to determine whether the clutch is in an engaged state.

[0026] In some embodiments, the preprocessing includes: image enhancement, binarization and morphological filtering to reduce noise interference.

[0027] In some embodiments, the binarization adopts Gaussian adaptive threshold method; let the pixel with coordinate position (x, y) of the original image I be I(x, y), and the area size be N w ×N h , the threshold T(x,y) is generated as follows:

[0028]

[0029] Among them, C is an adjustable parameter.

[0030] The present invention also provides an electronic device, comprising multiple processors, a memory, and a computer program stored in the memory and executable on the processor, wherein when the multiple processors execute the computer program, the machine vision-based tooth clutch engagement detection method as described in any one of the above items is implemented.

[0031] The present invention also provides a storage medium having a computer program stored thereon, wherein when the computer program is executed, the method for detecting engagement of a tooth clutch based on machine vision as described in any one of the above is implemented.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present application provides a tooth-type clutch engagement detection method, device, and medium based on machine vision, which uses edge detection and straight line detection technology to accurately judge the clutch engagement state, and realizes automatic and real-time detection of the clutch engagement state. Compared with the traditional manual detection method, this method has the advantages of high detection efficiency, high accuracy, and high safety. It can quickly and accurately judge the working state of the clutch in the industrial field, improve the reliability and stability of the system, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 Schematic diagram of a flow chart of an embodiment of the present invention.

[0035] Figure 2 Schematic diagram of the detection structure of an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] This embodiment provides a tooth clutch engagement detection method based on machine vision, which can accurately determine the clutch engagement state and help improve the automation and intelligence level of industrial sites.

[0038] Specifically, the method includes:

[0039] Step S1, acquiring an image of a tooth gap area of ​​a dog clutch and performing preprocessing.

[0040] From the working principle of the clutch, we know that when the gap between the upper and lower wheels of the clutch, that is, the width between the end teeth, is smaller than the standard tooth width, it can be determined that it is in the engaged state. Therefore, to detect the engaged state of the tooth clutch, that is, whether the tooth clutch is engaged, the distance between its active plate and the driven armature can be detected.

[0041] like Figure 2 As shown in the on-site schematic diagram, the detection camera and the clutch have a fixed position relationship. Since the camera optical axis is perpendicular to the transmission shaft of the clutch, the camera is always aimed at the clutch tooth gap area, so the edges at both ends of the clutch tooth gap should be parallel in the camera.

[0042] During the shooting process, in order to ensure that the camera can clearly capture the area of ​​the clutch tooth gap for accurate engagement state detection, the optical axis of the detection camera is perpendicular to the clutch transmission shaft, and the camera is installed at a position perpendicular to the clutch transmission shaft so that the detection camera can face the clutch tooth gap area. Ensure that the upper and lower edges of the tooth gap are parallel in the camera's field of view, which is convenient for subsequent edge detection and straight line extraction. The distance between the detection camera and the clutch is fixed to ensure stable imaging and avoid image deviation caused by changes in the position of the detection camera. At the same time, the distance setting also needs to take into account the focal length and viewing angle of the detection camera to ensure that the tooth gap area can be fully and clearly presented in the image. This relative position relationship ensures that the camera can accurately capture the edge of the clutch tooth gap, providing clear and accurate image input for the detection of the machine vision algorithm.

[0043] Preprocessing includes image enhancement, binarization and morphological filtering to reduce noise interference. First, the area where the tooth gap is located is located using the template matching technology based on Normalized Cross Correlation (NCC) and selected. This template matching is based on the size of the correlation coefficient of the pixel. Let the template image be T, the original image be I, and the correlation coefficient ρ of the pixel (x, y) with coordinates be:

[0044]

[0045] In the formula is the pixel mean of the template and the neighborhood of the same size as the template in the original image with (x, y) as the origin. Let M and N be the width and height of the template. The expression is as follows:

[0046]

[0047] Secondly, the selected image is processed by Gaussian adaptive binarization and adaptive morphological filtering to reduce noise interference. Binarization converts the image into a black and white binary image by setting a grayscale threshold. The morphological method refers to the use of several structural elements as "probes" on the binary image to fill the image with effective elements. According to the selection of structural elements, it is often used to filter the image or extract edges. The Gaussian adaptive threshold method calculates the corresponding threshold based on the specific features of each small area on the image. Therefore, different thresholds are used in different areas of the same image, which can well reflect the grayscale features of each image area without being interfered by other areas.

[0048] Suppose the pixel with coordinate position (x, y) of the original image I is I(x, y), and the area size is N w ×N h , select an appropriate adjustable parameter C. Generally, the adjustable parameter represents a constant subtracted from the average or weighted sum of the neighborhood pixels, which is related to the pixel distribution of the image. In this application environment, it is generally 30-50. The formula for the threshold T(x,y) is as follows:

[0049]

[0050] Step S2: Use the improved Canny operator to perform edge detection on the binarized and preprocessed image to obtain the edge of the clutch gear gap.

[0051] The improved Canny operator is used to detect the edge of the binary image, which can more effectively detect the real edge of the clutch tooth gap and provide support for the accuracy and stability of image processing. The improved Canny operator will use two templates that are differentiated in the bevel direction to calculate the gradient of each pixel. Two differential templates are introduced in the bevel direction of each pixel to capture more edge information. This multi-directional gradient calculation method can enhance the response to complex edges and improve the detection ability of the clutch tooth gap edge.

[0052] The improved Canny operator compares the calculated gradient value with the gradient value in the basic direction, thereby selecting the pixel with the maximum gradient as the edge, which can more accurately retain the edge pixels and reduce the edge misdetection and pseudo-edge caused by noise or unclear gradient. In non-maximum suppression, the maximum value is selected by comparing the gradient size in the basic direction. Due to the improvement of gradient calculation and non-maximum suppression in the oblique direction, this method can more clearly extract the upper and lower edges of the clutch tooth gap, which is crucial for subsequent straight line detection and tooth gap width measurement, ensuring the accuracy of clutch engagement detection.

[0053] Step S3, using the Hough transform method to detect parallel straight lines in the edge image, wherein the parallel straight lines are used to represent the upper and lower edges of the tooth gap;

[0054] Specifically, the parallel straight line edges of the tooth gap are identified, and the parallel characteristics of the parallel straight lines are determined by calculating the angle and distance relationship between adjacent edge points.

[0055] Hough Transform (HT) is used to detect straight lines in edge images. In Hough Transform, any straight line can be represented by two parameters (ρ, θ), where ρ is the vertical distance from the origin to the line, and θ is the angle formed by the vertical line and the horizontal axis measured in the counterclockwise direction. Hough Transform will traverse all edge points in the image and calculate the parameters of the straight lines determined by them in pairs. Straight lines with the same parameters will be regarded as the same straight line. Each time they appear, the "score" corresponding to the parameter of the straight line will increase by one. After traversing all edge points in the image, Hough Transform will obtain an integral table with (ρ, θ) as the horizontal and vertical axes. All possible straight lines have corresponding "scores". The higher the "score", the more likely the straight line exists in the original image. Hough Transform can identify straight line structures that are more obvious in edge images. For the detection of tooth clutch engagement, it can help detect the upper and lower edge straight lines of the clutch tooth gap, providing a basis for subsequent distance measurement.

[0056] Step S4, screening out parallel straight line pairs by using a mean shift clustering method, calculating the pixel width of the parallel straight line pairs, and converting the pixel width into an actual tooth gap width;

[0057] In order to screen out the straight line clusters with parallel relationships, the mean shift clustering method (MSC) is used. This is a non-parametric density-based clustering method that achieves clustering by continuously updating the center position of the data points. It does not need to pre-set the number of clusters, but determines the number of clusters by determining the number and position of the cluster center points. Therefore, it has excellent robustness to noise data. It is combined with the random sampling process of the straight line as the HT-MSC method.

[0058] Specifically, the mean shift clustering method is used to screen out parallel straight line pairs, including:

[0059] Step S41, determining the weight parameters and mean shift function of the parallel straight lines, and calculating the weighted mean shift vector of each data point;

[0060] Step S42, updating the data points until the weighted mean drift vectors of all data points converge to a preset value, and determining the cluster center;

[0061] Step S43: Determine a pair of parallel straight lines according to the cluster centers.

[0062] Considering that for two straight lines (ρ1, θ1) and (ρ2, θ2), combined with the geometric meaning of the parameters: if θ1 = θ2 (ρ1 ≠ ρ2), then the two straight lines are parallel. Therefore, we can first cluster the parameter points in a single dimension of parameter θ, and then obtain a set of parallel line clusters based on the peak points after clustering. Assume that a series of straight line parameter points have been obtained through the HT algorithm: (ρ i ,θ i ,S i ), where S i is the “score” of the line, and the “weight” of each parameter is:

[0063]

[0064] Among them, i is the parallel line whose weight parameter is being calculated, n is all the parallel lines detected, and j is the parallel lines that meet the condition L i Parallel straight lines; S j is the score of the parallel line j, ρ is the perpendicular distance from the origin to the parallel line, θ is the angle of the parallel line n, S n is the score of the parallel line n, L i is an angle θ i is a cluster of straight lines under the condition, l j is the set L i An element in .

[0065] This “weight” weighs the sum of all the straight line parameters θ equal to θ iThe cumulative score of the straight line will be used to provide an effective weight for the drift mean. According to the above score accumulation process, the effective parameter set Θ = {θ1, θ2, …, θ m}, the corresponding weight column is w={w1,w2,…,w m}, the weighted mean shift process is as follows:

[0066]

[0067] Among them, the weight is w={w1,w2,…,w m}, In the n+1th iteration, the angular center of the kth line cluster is updated; is the current angular center of the kth line cluster in the nth iteration, is the angle parameter of the i-th line in the n-th iteration; h is the bandwidth parameter, and g(x) is the kernel function.

[0068] Through the iterative update process of the above mean shift, the angle center is gradually Move in the direction of the maximum local density until the weighted mean drift vector of all data points converges to the preset value and determines the cluster center.

[0069] Since the screened parallel straight line pairs are parallel, we only need to calculate the perpendicular distance between the two straight lines to get the pixel width between them. In order to convert the pixel width into the actual tooth gap width, we need to use the camera calibration parameters, that is, the camera resolution and distance calibration. Assuming that the actual length corresponding to each pixel is d mm / pixel, the actual tooth gap width W can be calculated by the following formula:

[0070] W = pixel width × d

[0071] Wherein, d is the actual length corresponding to each pixel obtained by camera calibration (for example, 1 pixel = 0.1 mm).

[0072] Step S5: determining the engaged state of the dog clutch according to the actual tooth gap width.

[0073] The actual tooth gap width is compared with the preset engagement threshold to determine whether the clutch is in the engaged state. A preset engagement threshold, that is, the critical value of the clutch tooth gap width, needs to be set according to the design and working requirements of the tooth clutch. When the actual tooth gap width is less than or equal to the threshold, it indicates that the tooth clutch is in the engaged state; when the actual tooth gap width is greater than the threshold, it indicates that the tooth clutch is not engaged. This method realizes real-time monitoring and automatic judgment of the clutch to ensure the reliable operation of the mechanical system. By setting a suitable preset engagement threshold, the state of tooth clutches of different types and working requirements can be judged to adapt to a variety of industrial scenarios.

[0074] In order to experimentally verify the effectiveness of the engagement detection, a sample library consisting of 200 clutch images taken from different angles and containing two states is used to verify the accuracy of the detection scheme. The sample contains 100 engagement images and 100 non-engagement images. False detection means that the engagement is detected as non-engagement or the non-engagement is detected as engagement. Loss means that the system does not capture the clutch edge in the image and thus has no detection result. The detection results are shown in the following table:

[0075]

[0076] After verification, this method shows high accuracy under different states, among which the accuracy of the engaged state is slightly higher than that of the unengaged state.

[0077] The present application provides a tooth-type clutch engagement detection method, device, and medium based on machine vision, which uses edge detection and straight line detection technology to accurately judge the clutch engagement state, and realizes automatic and real-time detection of the clutch engagement state. Compared with the traditional manual detection method, this method has the advantages of high detection efficiency, high accuracy, and high safety. It can quickly and accurately judge the working state of the clutch in the industrial field, improve the reliability and stability of the system, and has broad application prospects.

[0078] Based on the same inventive concept, the present invention also provides a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the aforementioned method for detecting engagement of a tooth clutch based on machine vision are implemented.

[0079] Among them, the computer program includes computer program code, which can be in source code form, object code form, executable file or some intermediate form, etc. Computer readable media may include: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in computer readable media can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer readable media do not include electric carrier signals and telecommunication signals.

[0080] The present invention can be used in many general or special computing system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

[0081] The present application also provides an electronic device, comprising a memory, multiple processors, and a program stored in the memory, wherein the program is configured to be executed by the processors, and the steps of the above method are implemented when the multiple processors execute the program.

[0082] In addition, the present invention also provides a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the aforementioned method are implemented. The present invention can be used in many general or special computing system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc.

[0083] The device in this embodiment and the method in the aforementioned embodiment are two aspects based on the same inventive concept. The method implementation process has been described in detail above, so those skilled in the art can clearly understand the structure and implementation process of the system in this embodiment based on the aforementioned description. For the sake of brevity of the specification, they will not be repeated here.

[0084] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for detecting engagement of a tooth clutch based on machine vision, characterized in that: The method comprises the following steps: Acquire an image of the tooth gap area of ​​the dog clutch and perform preprocessing; The edge of the clutch gear gap is obtained by using the improved Canny operator to detect the edge of the preprocessed image. The Hough transform method is used to detect parallel straight lines in the edge image, and the parallel straight lines are used to represent the upper and lower edges of the tooth gap; Screening out parallel straight line pairs by using a mean shift clustering method, calculating the pixel width of the parallel straight line pairs, and converting the pixel width into an actual tooth gap width; The engagement state of the dog clutch is determined according to the actual tooth gap width.

2. The method for detecting engagement of a dog clutch based on machine vision according to claim 1, characterized in that: The mean shift clustering method is used to select parallel straight line pairs, including: Determine the weight parameters and mean shift function of the parallel straight lines, and calculate the weighted mean shift vector of each data point; Update the data points until the weighted mean drift vector of all data points converges to the preset value and determine the cluster center; A pair of parallel straight lines is determined according to the cluster centers.

3. The method for detecting engagement of a dog clutch based on machine vision according to claim 2, characterized in that: The weight parameters of the parallel straight lines are: Among them, i is the parallel line whose weight parameter is being calculated, n is all the parallel lines detected, and j is the parallel lines that meet the condition L i Parallel straight lines; S j is the score of the parallel line j, ρ is the perpendicular distance from the origin to the parallel line, θ is the angle of the parallel line n, S n is the score of the parallel line n, L i is an angle θ i is a cluster of straight lines under the condition, l j is the set L i An element in .

4. The method for detecting engagement of a dog clutch based on machine vision according to claim 3, characterized in that: The weighted mean shift vector for each data point is calculated using the following formula: Among them, the weight is w={w1,w2,…,w m }, In the n+1th iteration, the angular center of the kth line cluster is updated; is the current angular center of the kth line cluster in the nth iteration, is the angle parameter of the i-th line in the n-th iteration; h is the bandwidth parameter, and g(x) is the kernel function.

5. The method for detecting engagement of a dog clutch based on machine vision according to claim 1, characterized in that: The Hough transform method is used to detect parallel lines in edge images, including: Identify the parallel straight line edges of the tooth gap and determine the parallel characteristics of the parallel straight lines by calculating the angle and distance relationship between adjacent edge points.

6. The method for detecting engagement of a dog clutch based on machine vision according to claim 5, characterized in that: Determining the engagement state of the dog clutch according to the actual tooth gap width includes: The actual backlash width is compared with a preset engagement threshold to determine whether the clutch is in an engaged state.

7. The method for detecting engagement of a dog clutch based on machine vision according to claim 6, characterized in that: The preprocessing includes: image enhancement, binarization and morphological filtering to reduce noise interference.

8. The method for detecting engagement of a dog clutch based on machine vision according to claim 7, characterized in that: The binarization adopts Gaussian adaptive threshold method; let the pixel with coordinate position (x, y) of the original image I be I(x, y), and the area size be N w ×N h , the threshold T(x,y) is generated as follows: Among them, C is an adjustable parameter.

9. An electronic device, characterized in that: The method comprises a plurality of processors, a memory and a computer program stored in the memory and executable on the processor. When the plurality of processors execute the computer program, the method for detecting engagement of a tooth clutch based on machine vision as claimed in any one of claims 1 to 8 is implemented.

10. A storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed, the method for detecting engagement of a tooth clutch based on machine vision as claimed in any one of claims 1 to 8 is implemented.

Images