| As a non-contact non-destructive detecting technology,machine vision has the characteristics of high detection accuracy,fast processing speed,long continuous working time and strong stability.It has been widely used in the field of intelligent production.Machine vision technology uses industrial cameras to collect product images for analysis and processing to achieve product detection and identification automatically,thereby replacing tedious manual inspections.In recent years,visual measurement and visual inspection technologies based on machine vision have become more and more popular in the manufacturing industry.Aiming at the inspection requirements of the key features of the valve plate parts of automobile air-conditioning compressors,this research proposes a visual inspection system based on machine vision.The system detects and measures the overall structural defects,edge contour defects,hole spacing of parts.The main research contents of this paper are as follows:(1)According to the actual inspection needs,the research and design of the visual inspection system for the key parameters of the small valve plate has been completed.In terms of hardware,the hardware selection of the visual inspection system and the construction of the experimental platform are carried out.In terms of software,a visual inspection software system is designed and developed.(2)The camera calibration and image preprocessing algorithms are studied.The coordinate system conversion relationship of the ideal imaging model is analyzed,and the lens distortion in real imaging is considered for distortion correction.Use Zhang’s calibration method to complete the camera calibration.Analyze the collected images and perform image preprocessing according to the inspection requirements.(3)The inspection method of the overall structural integrity of the part is studied.In this paper,the number of through holes of the part is detected on the down-sampled image.By comparing the area with the standard value,structural defects can be judged,and parts with stamping defects can be quickly eliminated.(4)The detection method of part of the hole distance and aperture is studied.Combined with the Zernike moment sub-pixel detection method and the least square circle fitting,the sub-pixel level circle detection is realized.The ROI region is extracted from the preprocessed image by using the ROI region parameter to complete the sub-pixel level hole pitch measurement of the key holes.The Canny edge detection algorithm,the Zernike moment sub-pixel algorithm and the improved Zernike moment sub-pixel detection algorithm are compared and analyzed through experiments.(5)The detection method of the edge contour defect of the part is researched.Aiming at the inherent characteristic relationship of porous parts,a universal fast template matching algorithm based on porous centroids is proposed.Using the porous parameters of the part,the part can be quickly matched and compared with the standard template to complete the contour defect detection of the part.Based on the above research results,the software and hardware construction of the detection system has been completed,and the system experimental test has been carried out.Experiments have proved that the detection and measurement technology of the key parameter detection system has certain practical significance. |
