Research On Visual Detection System For Quality Of Upper Valve Spring Retainer Of Vehicle

As an important part of valve group support and fixing valve spring in automobile engine,the quality of upper valve spring retainer affects the working efficiency of automobile engine.With the proposal of "Made in China 2025",modern manufacturing industry has developed rapidly,and on the premise of maintaining product production efficiency,the focus of manufacturing industry has slowly shifted to improving product quality.The traditional manual sampling inspection is still used in the production process of the upper valve spring retainer.In the long-time work,it not only consumes a lot of manpower cost,but also the detection effect decreases rapidly,which cannot meet the increasing quality detection requirements.In order to solve the above problems,this thesis studies and designs a set of visual detection system for the quality of upper valve spring retainer.The main contents are as follows:(1)Aiming at the actual detection requirements,the software and hardware of the quality detection system are selected and designed to realize image acquisition and sorting.(2)Aiming at the problems of digital image preprocessing,the effects of different processing algorithms are analyzed,and finally,Fourier transform Gaussian low-pass filter is used to denoise,and linear gray transform is used to enhance the image.For the contour detection and fitting of inner and outer diameter circle in size detection,through comparative experiments,RANSAC pre fitting and least square circle fitting are used to realize size detection.(3)Aiming at the inwall scratch detection of the upper valve spring retainer,polar coordinate transformation,interpolation processing and dynamic double threshold local image enhancement processing are used to detect the accurate direction and length of the inwall scratch,which is convenient for traceability.Aiming at the problem of surface defect detection and classification,the similarities and differences of different geometric features of surface defects are analyzed,and multiple sets of thresholds are set for classification.(4)The experimental platform was built,and the upper computer interactive interface was built with Lab VIEW to verify the stability and accuracy of the visual detection algorithm for the quality of upper valve spring retainer.The experimental results show that the visual detection system for the quality of upper valve spring retainer has high stability and accuracy in the actual detection process,and the dimensional detection error is less than 0.045 mm,the accuracy of surface defect detection is more than 96.67%,and the accuracy of defect classification is more than 95.00%.The indicators meet the expectations,and have practical application value.The detection system can be further improved and put it into the actual production process.