| Non-standard parts of automobile is an important part of automobile.Defects of nonstandard parts are an important factor affecting the quality of automobile.It is also the key to realize intelligent manufacturing in automobile manufacturing industry.At present,the detection methods of automobile non-standard products are mainly manual detection,which results in low detection efficiency,high error rate and high labor intensity,resulting in high product cost.For this reason,aiming at the problem of automatic detection of automotive engine heat shield for non-standard products,this dissertation carries out in-depth research based on machine vision technology,mainly studies the existence of rivets for heat shield and the development and implementation of detection methods,detection technology and detection system for concave and convex identification.The main research work includes the following four aspects:(1)Systematically elaborated the automobile non-standard spare parts product automation inspection demand and the present situation.(2)The detection system of automobile non-standard parts was established,which mainly includes the optimization and selection of camera,lens,light source and lighting system.The system meets the requirements of visual inspection of non-standard parts.(3)In the aspect of noise removal,by comparing the effect maps of four methods of image filtering,it is found that the adaptive median filtering method has the best effect in preserving image edge information,and the details of edge information are more abundant.(4)The detection algorithm of non-standard automotive products is discussed in detail.Taking the detection of non-standard automotive parts of heat shield as an example,the detection method of non-standard automotive products is elaborated.The detection system of non-standard automotive products is developed based on HALCON and C++ hybrid programming technology. |