Circular run-out detection method and system based on machine vision

Through machine vision-based detection methods, the circular jump feature points of the vehicle blower impeller are extracted using CCD industrial cameras and image processing technology, and the problems of low detection efficiency and low accuracy in the existing technology are solved, and efficient and accurate circles are achieved Beat detection.

CN118967629AInactive Publication Date: 2024-11-15YANCHENG INST OF TECH
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411059623.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-05
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing automotive blower impeller circular jump detection technology is inefficient and has low accuracy, especially in batch inspection and high accuracy requirements, and the automatic measurement equipment is expensive and cumbersome to install and debug.

Method used

Using machine vision-based detection method, the CCD industrial camera is used to obtain the video of the impeller of the car blower around the axis, and the target contour intersection point and Shi-Tomas i corner point feature points are extracted through image processing, and the circular jump tolerance value is calculated, combining light source illumination and Image calibration to realize automated detection.

Benefits of technology

It improves the efficiency and accuracy of circular jump detection of automobile blower impeller, reduces the burden of software programming, realizes real-time online feedback and large-scale repetitive detection, and avoids the introduction of human-factor errors.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118967629A_ABST
    Figure CN118967629A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of automobile air blower impeller detection, and provides a circle run-out detection method and system based on machine vision, and the method comprises the steps: obtaining a video that an automobile air conditioner air blower impeller rotates around an axis; processing the video to obtain a target area image; extracting a target contour intersection point in the target area image, and obtaining a floating range of the target contour intersection point after rotation; extracting Shii-Toma i angular point feature points in the target area image, and obtaining a jumping range of the Shii-Toma i angular point feature points; and calculating to obtain a circular run-out tolerance value based on the floating range and the run-out range. According to the invention, the CCD industrial camera is adopted to carry out automobile blower impeller circle run-out detection, so that the burden of software programming is reduced, the requirement of automobile blower impeller circle run-out detection is met, and the efficiency and accuracy of automobile blower impeller circle run-out detection are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of automobile blower impeller detection, and in particular to a circular runout detection method and system based on machine vision. Background Art

[0002] With the vigorous development of intelligence in new energy vehicles, the quality inspection of its components also needs to keep pace with the times. The circular runout inspection of the blower impeller is an indispensable part of the blower quality inspection. The circular runout error is due to various factors such as uneven material or blank defects, errors in processing and assembly, and even asymmetric geometric shapes in design, which cause the imbalance of the rotating body. When the circular runout error is unqualified, it will generate vibration and noise during rotation, accelerate bearing wear, shorten the life of the machine, and even cause destructive accidents in severe cases.

[0003] The circular runout detection of the impeller of the automobile air-conditioning blower refers to the measurement of the peak-to-peak value of the upper cylindrical surface when the impeller rotates around the axis for one circle, and the difference between the maximum and minimum readings measured by the fixed indicator; the runout tolerance is the geometric tolerance of controlling the runout change value of the target element by setting a straight line as the rotation axis and rotating the target object; the detection methods are mainly divided into contact and non-contact;

[0004] Contact measurement is divided into manual inspection using measuring instruments, such as V-bracket measurement and deflection meter measurement, and automatic measurement using precision instruments, such as three-coordinate measuring machines and roundness meters. Manual inspection is limited by the experience of the inspector, has low efficiency, and is not suitable for batch inspection. Automatic measurement equipment is expensive and the installation and debugging process is cumbersome.

[0005] Non-contact measurement mainly includes laser measurement and image processing methods, such as measurement based on laser triangulation principle and microscope; laser displacement sensor measurement, the measurement accuracy depends on the accuracy of the installation of the laser sensor, and is not used to measure the total runout error; high-precision CCD micrometer measurement requires the transmitter and receiver to be installed horizontally and vertically.

[0006] The machine vision-based method uses a camera to capture real images, obtains effective information through image processing to complete the task of controlling the machine, and uses the camera to simulate the human eye to complete the measurement and analysis process. It has the advantages of real-time online feedback, avoiding the introduction of human errors, and realizing large-scale repeatable detection to improve measurement accuracy and detection speed.

[0007] Therefore, it is necessary to provide a circular runout detection method and system based on machine vision. Summary of the invention

[0008] The present invention provides a circular runout detection method and system based on machine vision, which adopts a CCD industrial camera to perform circular runout detection on an automobile blower impeller, thereby reducing the burden of software programming, meeting the circular runout detection requirements of an automobile blower impeller, and improving the efficiency and accuracy of circular runout detection on an automobile blower impeller.

[0009] The present invention provides a circular runout detection method based on machine vision, comprising:

[0010] Get a video of the car air conditioner blower wheel rotating around its axis;

[0011] Process the video to obtain the target area image;

[0012] Extract the target contour intersection point in the target area image, and obtain the floating range of the target contour intersection point after rotation;

[0013] Extract Shi-Tomas i corner feature points in the target area image and obtain the jump range of Shi-Tomas i corner feature points;

[0014] Based on the floating range and run-out range, the circular run-out tolerance value is calculated.

[0015] Furthermore, a video of the automobile air-conditioning blower impeller rotating around the axis is obtained, including:

[0016] Configure the CCD industrial camera as a camera device;

[0017] Adjust the parameters of the CCD industrial camera to obtain the camera parameters;

[0018] Configure the light source lighting scene as backlighting;

[0019] Using a CCD industrial camera, according to the camera parameters, a video of the automobile air-conditioning blower impeller rotating around its axis is obtained in a backlight illumination scene.

[0020] Furthermore, before obtaining a video of the impeller of the automobile air-conditioning blower rotating around the axis in the backlight illumination scene using the CCD industrial camera according to the camera parameters, the CCD industrial camera is calibrated, specifically:

[0021] Based on the set chessboard calibration plate, 20 chessboard images at different angles are taken using a CCD industrial camera.

[0022] Get the corner coordinates of each chessboard image;

[0023] According to the coordinates of the corner points, the calibration function cal ibrateCamera in the computer vision library Opencv is used to calculate the parameter matrix and distortion coefficient of the CCD industrial camera;

[0024] For the parameter matrix and distortion coefficient, the undistort function in the computer vision library Opencv is used to dedistort and complete the calibration of the CCD industrial camera.

[0025] Furthermore, the video is processed to obtain an image of the target area, including:

[0026] Extract the side image of the automobile air-conditioning blower wheel rotating around the axis in the video;

[0027] The side image is processed using an image processing algorithm to obtain an image of the target area.

[0028] Furthermore, the image processing algorithm includes:

[0029] The side image is grayed out by using a weighted average to obtain a grayed out image;

[0030] The median filtering algorithm is used to filter the grayscale processed image to obtain a filtered processed image;

[0031] The global threshold processing method is used to sharpen the edges of the filtered image to obtain an enhanced image.

[0032] The Canny operator is used to perform edge detection on the enhanced image to obtain the target area image; in the denoising step in the edge detection process, the median filter is used to smooth the image to better retain the edge details; in the step of applying double thresholds to determine possible boundaries in the edge detection process, the double thresholds are manually tested to obtain more suitable double threshold values.

[0033] Furthermore, the target contour intersection in the target area image is extracted, and the floating range of the target contour intersection after rotation is obtained, including:

[0034] Extracting contour features in the target area image to obtain a first contour;

[0035] After using the bicubic linear interpolation method to interpolate the contour in the target area image, the contour feature is extracted to obtain the second contour;

[0036] Superimposing the first contour and the second contour to obtain a target contour;

[0037] Get the intersection points of the target contour with the horizontal coordinate axis and the vertical coordinate axis after rotation, and get the floating range of the intersection points.

[0038] Further, extract the Shi-Tomas i corner feature points in the target area image, and obtain the jump range of the Shi-Tomas i corner feature points, including:

[0039] The Lucas-Kanade optical flow method is used to track the coordinates of the corner feature points to describe the motion of the blower impeller and obtain the beating range of the corner feature points; the accuracy of the coordinates is set to 0.3.

[0040] Furthermore, it also includes: storing and managing the floating range, the run-out range and the circular run-out tolerance value, specifically:

[0041] Build a data repository and use the data storage template in the data repository to store the floating range, runout range and circular runout tolerance values;

[0042] Based on the data filtering rules set in the data storage template, the obtained floating range and the jumping range are filtered to obtain the first floating range and the first jumping range that meet the detection requirements;

[0043] Based on a display interface set on a computer terminal or a data display program on a mobile terminal APP, the first floating range, the first run-out range and the circular run-out tolerance value are displayed;

[0044] Based on the data calling template set in the data storage library, the first floating range, the first run-out range and the circular run-out tolerance value are called for detection and analysis of the circular run-out detection result.

[0045] Furthermore, it also includes adjusting the circular runout detection parameters according to the detection accuracy requirements based on the historical data in the data repository; specifically:

[0046] Based on the data repository, the historical first floating range, the historical first runout range and the historical circle runout tolerance value, as well as the corresponding historical detection accuracy, are obtained;

[0047] Get the historical camera parameters of CCD industrial cameras;

[0048] Based on historical camera parameters and the historical first floating range, the historical first run-out range and the historical circle run-out tolerance value, historical detection accuracy, and historical camera parameters, a knowledge graph structure network model is constructed with detection accuracy-camera parameters-detection results as entities, where the detection results are the historical first floating range, the historical first run-out range and the historical circle run-out tolerance value;

[0049] Obtain the detection accuracy of circular runout to be achieved, and use the knowledge graph structure network model to recommend the camera parameters to be adjusted based on the detection accuracy of circular runout to be achieved;

[0050] Based on the video parameters to be adjusted, the parameters of the CCD industrial camera are adjusted to capture the video of the automobile air-conditioning blower impeller rotating around the axis.

[0051] The circular runout detection system based on machine vision includes:

[0052] A video acquisition unit, used to acquire a video of the automobile air-conditioning blower impeller rotating around the axis;

[0053] A video processing unit, used for processing the video to obtain an image of the target area;

[0054] A target contour intersection floating range acquisition unit is used to extract the target contour intersection in the target area image and acquire the floating range of the target contour intersection after rotation;

[0055] A corner feature point jump range acquisition unit is used to extract Shi-Tomas i corner feature points in the target area image and acquire the jump range of Shi-Tomas i corner feature points;

[0056] The circular runout tolerance value acquisition unit is used to calculate and obtain the circular runout tolerance value based on the floating range and the runout range.

[0057] Compared with the prior art, the present invention has the following advantages and beneficial effects: using a CCD industrial camera to detect the circular runout of an automobile blower impeller is conducive to reducing the burden of software programming, meeting the requirements for the circular runout detection of an automobile blower impeller, and improving the efficiency and accuracy of the circular runout detection of an automobile blower impeller.

[0058] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0059] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0061] Figure 1 Schematic diagram of the steps of the circular runout detection method based on machine vision;

[0062] Figure 2 A schematic diagram of the steps of a method for processing a video and obtaining an image of a target area;

[0063] Figure 3 This is a structural diagram of the circular runout detection system based on machine vision. DETAILED DESCRIPTION

[0064] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0065] The present invention provides a circular runout detection method based on machine vision, such as Figure 1 As shown, including:

[0066] Get a video of the car air conditioner blower wheel rotating around its axis;

[0067] Process the video to obtain the target area image;

[0068] Extract the target contour intersection point in the target area image, and obtain the floating range of the target contour intersection point after rotation;

[0069] Extract Shi-Tomas i corner feature points in the target area image and obtain the jump range of Shi-Tomas i corner feature points;

[0070] Based on the floating range and run-out range, the circular run-out tolerance value is calculated.

[0071] The working principle of the above technical solution is: in order to realize the circular runout detection method based on machine vision, the present invention first obtains a video of the automobile air-conditioning blower impeller rotating around the axis; processes the video to obtain a target area image; then extracts the target contour intersection in the target area image, and obtains the floating range of the target contour intersection after rotation; extracts the Shi-Tomas i corner feature points in the target area image, and obtains the runout range of the Shi-Tomas i corner feature points; finally, based on the floating range and the runout range, the circular runout tolerance value is calculated.

[0072] The beneficial effects of the above technical solution are: adopting the solution provided in this embodiment and using a CCD industrial camera to perform circular runout detection of automobile blower impellers can help reduce the burden of software programming, meet the requirements for circular runout detection of automobile blower impellers, and improve the efficiency and accuracy of circular runout detection of automobile blower impellers.

[0073] In one embodiment, obtaining a video of an automobile air conditioner blower impeller rotating around an axis includes:

[0074] Configure the CCD industrial camera as a camera device;

[0075] Adjust the parameters of the CCD industrial camera to obtain the imaging parameters; the imaging parameters include: resolution of 1280*720, pixel size of 3.75um*3.75um, target size of 1 / 3", and fixed lens focal length of 3.4mm;

[0076] Configure the light source lighting scene as backlighting;

[0077] Using a CCD industrial camera, according to the camera parameters, a video of the automobile air-conditioning blower impeller rotating around its axis is obtained in a backlight illumination scene.

[0078] The working principle of the above technical solution is: obtaining a video of the impeller of an automobile air-conditioning blower rotating around the axis, including: configuring a CCD industrial camera as a camera device; adjusting the parameters of the CCD industrial camera to obtain camera parameters; the camera parameters include: a resolution of 1280*720, a pixel size of 3.75 um*3.75um, a target surface size of 1 / 3", and a fixed lens focal length of 3.4mm; configuring the light source illumination scene to be backlight illumination; using the CCD industrial camera, according to the camera parameters, in the light source illumination scene of backlight illumination, obtaining a video of the impeller of an automobile air-conditioning blower rotating around the axis.

[0079] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, according to the camera parameters, in the backlight lighting scene, an accurate video of the automobile air-conditioning blower impeller rotating around the axis can be obtained.

[0080] In one embodiment, before obtaining a video of the impeller of the automobile air-conditioning blower rotating around the axis in a backlight illumination scene using a CCD industrial camera according to the camera parameters, the CCD industrial camera is calibrated, specifically:

[0081] Based on the set chessboard calibration plate, 20 chessboard images at different angles are taken using a CCD industrial camera.

[0082] Get the corner coordinates of each chessboard image;

[0083] According to the coordinates of the corner points, the calibration function cal ibrateCamera in the computer vision library Opencv is used to calculate the parameter matrix and distortion coefficient of the CCD industrial camera;

[0084] For the parameter matrix and distortion coefficient, the undistort function in the computer vision library Opencv is used to dedistort and complete the calibration of the CCD industrial camera.

[0085] The working principle of the above technical solution is: using a CCD industrial camera, according to the camera parameters, in the backlight illumination scene, before obtaining the video of the automobile air-conditioning blower impeller rotating around the axis, it also includes calibrating the CCD industrial camera, first based on the set chessboard calibration plate, using the CCD industrial camera to shoot 20 chessboard pictures at different angles; obtaining the corner point coordinates of each chessboard picture; according to the corner point coordinates, using the calibration function cal ibrateCamera in the computer vision library Opencv, calculate the parameter matrix and distortion coefficient of the CCD industrial camera; for the parameter matrix and distortion coefficient, using the undistort function in the computer vision library Opencv to remove the distortion, and complete the calibration of the CCD industrial camera.

[0086] The beneficial effect of the above technical solution is: by adopting the solution provided by this embodiment, camera calibration is to solve the problem of image distortion, which is conducive to improving the camera accuracy.

[0087] In one embodiment, Figure 2 As shown, the video is processed to obtain the target area image, including:

[0088] Extract the side image of the automobile air-conditioning blower wheel rotating around the axis in the video;

[0089] The side image is processed using an image processing algorithm to obtain an image of the target area.

[0090] The working principle of the above technical solution is: in order to obtain the target area image, the present invention extracts the side image of the automobile air-conditioning blower impeller rotating around the axis in the video; uses the image processing algorithm to process the side image to obtain the target area image.

[0091] The beneficial effect of the above technical solution is: by adopting the solution provided by this embodiment, by obtaining the target area image, basic conditions are provided for subsequent processing.

[0092] In one embodiment, the image processing algorithm includes:

[0093] The side image is grayed out by using a weighted average to obtain a grayed out image;

[0094] The median filtering algorithm is used to filter the grayscale processed image to obtain a filtered processed image;

[0095] The global threshold processing method is used to sharpen the edges of the filtered image to obtain an enhanced image.

[0096] The Canny operator is used to perform edge detection on the enhanced image to obtain the target area image; in the denoising step in the edge detection process, the median filter is used to smooth the image to better retain the edge details; in the step of applying double thresholds to determine possible boundaries in the edge detection process, the double thresholds are manually tested to obtain more suitable double threshold values.

[0097] The working principle of the above technical solution is as follows: grayscale processing can better retain the details of the image, and can avoid the image distortion problems caused by the average grayscale method and the maximum grayscale method; median filtering overcomes the problem of blurred details that are prone to occur in the mean and Gaussian filtering algorithms, and can effectively protect the edge, texture and other feature information of the target; since contour search requires the background to be black and the target to be white, the threshold segmentation inversion operation is selected. After the threshold segmentation processing, the outline of the impeller is clearer, reducing the subsequent image processing time and improving efficiency; image enhancement can highlight the edge of the impeller, so the image is threshold segmented. Since the object to be measured is completely different from the background, the global threshold processing method is used to sharpen the edge.

[0098] The beneficial effect of the above technical solution is: by adopting the solution provided by this embodiment, an image that meets the requirements can be obtained by using an image processing algorithm.

[0099] In one embodiment, extracting the target contour intersection point in the target area image and obtaining the floating range of the target contour intersection point after rotation includes:

[0100] Extracting contour features in the target area image to obtain a first contour;

[0101] After using the bicubic linear interpolation method to interpolate the contour in the target area image, the contour feature is extracted to obtain the second contour;

[0102] Superimposing the first contour and the second contour to obtain a target contour;

[0103] Get the intersection points of the target contour with the horizontal coordinate axis and the vertical coordinate axis after rotation, and get the floating range of the intersection points.

[0104] The working principle of the above technical solution is: in order to extract the intersection of the target contour in the target area image and obtain the floating range of the intersection of the target contour after rotation, the present invention extracts the contour features in the target area image to obtain a first contour; then uses the bicubic linear interpolation method to interpolate the contour in the target area image, and then extracts the contour features to obtain a second contour; superimposes the first contour and the second contour to obtain the target contour; respectively obtains the intersection of the target contour and the horizontal coordinate axis and the vertical coordinate axis after rotation, and obtains the floating range of the intersection.

[0105] The beneficial effect of the above technical solution is: in order to obtain a more accurate edge contour, the solution provided by this embodiment is adopted, and in view of the discontinuity of contour points, bicubic linear interpolation is used to interpolate the contour, thereby increasing the number of contour points and subdividing the contour.

[0106] In one embodiment, extracting Shi-Tomas i corner feature points in the target area image and obtaining the jitter range of the Shi-Tomas i corner feature points includes:

[0107] The Lucas-Kanade optical flow method is used to track the coordinates of the corner feature points to describe the motion of the blower impeller and obtain the beating range of the corner feature points; the accuracy of the coordinates is set to 0.3.

[0108] The working principle of the above technical solution is: extracting Shi-Tomas i corner feature points in the target area image, and obtaining the runout range of the Shi-Tomas i corner feature points, including: using the Lucas-Kanade optical flow method to track the coordinates of the corner feature points to describe the movement of the blower impeller, and obtaining the runout range of the corner feature points; wherein the accuracy of the coordinates is set to 0.3; Shi-Tomas i corner detection is a feature point detection algorithm in computer vision, which can find pixels with significant changes in an image, and these pixels are usually called corner points; the Shi-Tomas i corner detection method is based on the Harris corner detection algorithm, but it is more robust to noise and grayscale changes and is more commonly used in practical applications; for the corner coordinates, the blower impeller is rotated three times to measure its rotation circle runout value.

[0109] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, the coordinates of the corner feature points are tracked by using the Lucas-Kanade optical flow method, thereby ensuring the accuracy of obtaining the jumping range of the corner feature points.

[0110] In one embodiment, the invention further includes: storing and managing the floating range, the run-out range and the circular run-out tolerance value, specifically:

[0111] Build a data repository and use the data storage template in the data repository to store the floating range, runout range and circular runout tolerance values;

[0112] Based on the data filtering rules set in the data storage template, the obtained floating range and the jumping range are filtered to obtain the first floating range and the first jumping range that meet the detection requirements;

[0113] Based on a display interface set on a computer terminal or a data display program on a mobile terminal APP, the first floating range, the first run-out range and the circular run-out tolerance value are displayed;

[0114] Based on the data calling template set in the data storage library, the first floating range, the first run-out range and the circular run-out tolerance value are called for detection and analysis of the circular run-out detection result.

[0115] The working principle of the above technical scheme is: in order to achieve effective management of detection data, the present invention constructs a data storage library, and uses the data storage template in the data storage library to store the floating range, run-out range and circular run-out tolerance value; based on the data filtering rules set in the data storage template, the obtained floating range and run-out range are filtered to obtain the first floating range and the first run-out range that meet the detection requirements; based on the display interface set on the computer terminal, or the data display program on the mobile terminal APP, the first floating range, the first run-out range and the circular run-out tolerance value are displayed; based on the data calling template set in the data storage library, the first floating range, the first run-out range and the circular run-out tolerance value are called for detection and analysis of the circular run-out detection results.

[0116] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, by storing, filtering, displaying and calling the floating range, run-out range and circular run-out tolerance value, effective management of the detection data can be achieved, which is conducive to the efficient analysis and utilization of the detection data.

[0117] In one embodiment, the method further includes adjusting the circular runout detection parameters based on the historical data in the data repository according to the detection accuracy requirements; specifically, the method includes:

[0118] Based on the data repository, the historical first floating range, the historical first runout range and the historical circle runout tolerance value, as well as the corresponding historical detection accuracy, are obtained;

[0119] Get the historical camera parameters of CCD industrial cameras;

[0120] Based on historical camera parameters and the historical first floating range, the historical first run-out range and the historical circle run-out tolerance value, historical detection accuracy, and historical camera parameters, a knowledge graph structure network model is constructed with detection accuracy-camera parameters-detection results as entities, where the detection results are the historical first floating range, the historical first run-out range and the historical circle run-out tolerance value;

[0121] Obtain the detection accuracy of circular runout to be achieved, and use the knowledge graph structure network model to recommend the camera parameters to be adjusted based on the detection accuracy of circular runout to be achieved;

[0122] Based on the video parameters to be adjusted, the parameters of the CCD industrial camera are adjusted to capture the video of the automobile air-conditioning blower impeller rotating around the axis.

[0123] The working principle of the above technical solution is as follows: in order to meet the detection requirements of different detection accuracies, the present invention adjusts the circular runout detection parameters according to the detection accuracy requirements based on the historical data in the data repository; first, based on the data repository, obtain the historical first floating range, the historical first runout range and the historical circular runout tolerance value, as well as the corresponding historical detection accuracy; obtain the historical camera parameters of the CCD industrial camera; then, based on the historical camera parameters and the historical first floating range, the historical first runout range and the historical circular runout tolerance value, the historical detection accuracy, and the historical camera parameters, use the knowledge graph structure to construct a knowledge graph structure network model in which the detection accuracy-camera parameters-detection results are entities, wherein the detection result is the historical first floating range, the historical first runout range and the historical circular runout tolerance value; finally, obtain the detection accuracy of the circular runout to be achieved, and according to the detection accuracy of the circular runout to be achieved, use the knowledge graph structure network model to recommend the camera parameters to be adjusted; based on the camera parameters to be adjusted, adjust the parameters of the CCD industrial camera to capture the video of the automobile air-conditioning blower impeller rotating around the axis.

[0124] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, by setting the knowledge graph structure network model, the parameters of the CCD industrial camera can be adjusted in a timely and effective manner to meet the adjustment of the camera parameters under different detection accuracy requirements.

[0125] Circular runout detection system based on machine vision, such as Figure 3 As shown, including:

[0126] A video acquisition unit, used to acquire a video of the automobile air-conditioning blower impeller rotating around the axis;

[0127] A video processing unit, used for processing the video to obtain an image of the target area;

[0128] A target contour intersection floating range acquisition unit is used to extract the target contour intersection in the target area image and acquire the floating range of the target contour intersection after rotation;

[0129] A corner feature point jump range acquisition unit is used to extract Shi-Tomas i corner feature points in the target area image and acquire the jump range of Shi-Tomas i corner feature points;

[0130] The circular runout tolerance value acquisition unit is used to calculate and obtain the circular runout tolerance value based on the floating range and the runout range.

[0131] The working principle of the above technical scheme is: in order to realize a circular runout detection system based on machine vision, the present invention proposes a video acquisition unit for acquiring a video of an automobile air-conditioning blower impeller rotating around an axis; proposes a video processing unit for processing the video to obtain a target area image; proposes a target contour intersection floating range acquisition unit for extracting the target contour intersection in the target area image, and obtaining the floating range of the target contour intersection after rotation; proposes a corner point feature point runout range acquisition unit for extracting the Shi-Tomas i corner point feature points in the target area image, and obtaining the runout range of the Shi-Tomas i corner point feature points; proposes a circular runout tolerance value acquisition unit for calculating the circular runout tolerance value based on the floating range and the runout range.

[0132] The beneficial effects of the above technical solution are: adopting the solution provided in this embodiment and using a CCD industrial camera to perform circular runout detection of automobile blower impellers can help reduce the burden of software programming, meet the requirements for circular runout detection of automobile blower impellers, and improve the efficiency and accuracy of circular runout detection of automobile blower impellers.

[0133] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A circular runout detection method based on machine vision, characterized in that: include: Get a video of the car air conditioner blower wheel rotating around its axis; Process the video to obtain the target area image; Extract the target contour intersection point in the target area image, and obtain the floating range of the target contour intersection point after rotation; Extract Shi-Tomasi corner feature points in the target area image and obtain the jump range of Shi-Tomasi corner feature points; Based on the floating range and run-out range, the circular run-out tolerance value is calculated.

2. The circular runout detection method based on machine vision according to claim 1, characterized in that: Get a video of the car air conditioner blower wheel rotating around its axis, including: Configure the CCD industrial camera as a camera device; Adjust the parameters of the CCD industrial camera to obtain the camera parameters; Configure the light source lighting scene as backlighting; Using a CCD industrial camera, according to the camera parameters, a video of the car air-conditioning blower impeller rotating around its axis is obtained in the backlight illumination scene.

3. The circular runout detection method based on machine vision according to claim 2, characterized in that: Before obtaining the video of the automobile air-conditioning blower impeller rotating around the axis in the backlight illumination scene using a CCD industrial camera according to the camera parameters, the CCD industrial camera is calibrated, specifically: Based on the set chessboard calibration plate, 20 chessboard images at different angles are taken using a CCD industrial camera. Get the corner coordinates of each chessboard image; According to the coordinates of the corner points, the calibration function calibrateCamera in the computer vision library Opencv is used to calculate the parameter matrix and distortion coefficient of the CCD industrial camera; For the parameter matrix and distortion coefficient, the undistort function in the computer vision library Opencv is used to dedistort and complete the calibration of the CCD industrial camera.

4. The circular runout detection method based on machine vision according to claim 1, characterized in that: Process the video to obtain the target area image, including: Extract the side image of the automobile air-conditioning blower wheel rotating around the axis in the video; The side image is processed using an image processing algorithm to obtain an image of the target area.

5. The circular runout detection method based on machine vision according to claim 4, characterized in that: Image processing algorithms include: The side image is grayed out by using a weighted average to obtain a grayed out image; The median filtering algorithm is used to filter the grayscale processed image to obtain a filtered processed image; The global threshold processing method is used to sharpen the edges of the filtered image to obtain an enhanced image. The Canny operator is used to perform edge detection on the enhanced image to obtain the target area image; in the denoising step in the edge detection process, the median filter is used to smooth the image to better retain the edge details; in the step of applying double thresholds to determine possible boundaries in the edge detection process, the double thresholds are manually tested to obtain more suitable double threshold values.

6. The circular runout detection method based on machine vision according to claim 1, characterized in that: Extract the target contour intersection in the target area image, and obtain the floating range of the target contour intersection after rotation, including: Extracting contour features in the target area image to obtain a first contour; After using the bicubic linear interpolation method to interpolate the contour in the target area image, the contour feature is extracted to obtain the second contour; Superimposing the first contour and the second contour to obtain a target contour; Get the intersection points of the target contour with the horizontal coordinate axis and the vertical coordinate axis after rotation, and get the floating range of the intersection points.

7. The circular runout detection method based on machine vision according to claim 1, characterized in that: Extract the Shi-Tomasi corner feature points in the target area image and obtain the jump range of the Shi-Tomasi corner feature points, including: The Lucas-Kanade optical flow method is used to track the coordinates of the corner feature points to describe the motion of the blower impeller and obtain the beating range of the corner feature points; the accuracy of the coordinates is set to 0.

3.

8. The circular runout detection method based on machine vision according to claim 1, characterized in that: Also includes: The data of the floating range, runout range and circular runout tolerance value are stored and managed, specifically: Build a data repository and use the data storage template in the data repository to store the floating range, runout range and circular runout tolerance values; Based on the data filtering rules set in the data storage template, the obtained floating range and the jumping range are filtered to obtain the first floating range and the first jumping range that meet the detection requirements; Based on a display interface set on a computer terminal or a data display program on a mobile terminal APP, the first floating range, the first run-out range and the circular run-out tolerance value are displayed; Based on the data calling template set in the data storage library, the first floating range, the first run-out range and the circular run-out tolerance value are called for detection and analysis of the circular run-out detection result.

9. The circular runout detection method based on machine vision according to claim 8, characterized in that: It also includes adjusting the circular runout detection parameters according to the detection accuracy requirements based on the historical data in the data repository; specifically: Based on the data repository, the historical first floating range, the historical first runout range and the historical circle runout tolerance value, as well as the corresponding historical detection accuracy, are obtained; Get the historical camera parameters of CCD industrial cameras; Based on historical camera parameters and the historical first floating range, the historical first run-out range and the historical circle run-out tolerance value, historical detection accuracy, and historical camera parameters, a knowledge graph structure network model is constructed with detection accuracy-camera parameters-detection results as entities, where the detection results are the historical first floating range, the historical first run-out range and the historical circle run-out tolerance value; Obtain the detection accuracy of circular runout to be achieved, and use the knowledge graph structure network model to recommend the camera parameters to be adjusted based on the detection accuracy of circular runout to be achieved; Based on the video parameters to be adjusted, the parameters of the CCD industrial camera are adjusted to capture the video of the automobile air-conditioning blower impeller rotating around the axis.

10. The circular runout detection system based on machine vision is characterized by: include: A video acquisition unit, used to acquire a video of the automobile air-conditioning blower impeller rotating around the axis; A video processing unit, used for processing the video to obtain an image of the target area; A target contour intersection floating range acquisition unit is used to extract the target contour intersection in the target area image and acquire the floating range of the target contour intersection after rotation; A corner feature point jump range acquisition unit is used to extract Shi-Tomasi corner feature points in the target area image and acquire the jump range of the Shi-Tomasi corner feature points; The circular runout tolerance value acquisition unit is used to calculate and obtain the circular runout tolerance value based on the floating range and the runout range.

Citation Information

Patent Citations

  • Contour detection method based on visual integrity and local feature fusion

    CN109558880A

  • Electrified railway insulator form and location tolerance detection device

    CN110514118A

  • Regional contour determination method and device and image display equipment

    CN110705576A

  • 360 degree optical measuring device

    CN112219085A

  • Blood vessel lumen intimal contour extraction method and device, ultrasonic equipment and storage medium

    CN114693710A