Research And Development Of Machine Vision Based Automotive Radar Bracket Assembly Quality Detection System And Key Technologies

The assembly quality of automobile radar support will affect the accuracy of radar detection and the safety performance of vehicle in normal use.At present,most enterprises use manual to detect radar assembly accuracy,the production efficiency is low,product quality is difficult to be guaranteed.In view of this,after the full research of the current industrial detection method based on machine vision technology,and in view of the imaging characteristics of automobile radar bracket assembly image,a method of assembly quality detection based on machine vision technology for radar bracket assembly is proposed.The assembly quality of automobile radar support can be detected automatically.In the process of research and development,this paper mainly includes the following aspects:(1)The whole design of automobile radar assembly quality inspection system.According to the characteristics of automobile radar support assembly,the overall design of vehicle radar support assembly quality detection system based on machine vision is carried out.The detection system includes three modules: incoming and outgoing module,detection module and welding module.According to the requirements of the detection system,the hardware design and selection are completed.A high performance industrial computer,four independent high resolution linear scanning cameras and a camera lens equipped with LED ring light source are adopted to complete the hardware platform of the detection system.(2)Radar assembly image preprocessing.According to the characteristics of automobile radar bracket assembly image,a gray world image enhancement method based on histogram is proposed.While the image of automobile radar bracket is enhanced by histogram equalization,the color balance is carried out by using gray world algorithm.Removing the influence of artificial light source on the image and enhancing the edge details of the image,three common filtering methods and principles are analyzed,and the filtering effects of different methods are studied.Finally,the image preprocessing based on median filter is determined.(3)Extraction of radar support profile.After studying the image segmentation method,an improved two-dimensional OTSU image segmentation method is proposed for the first time.The edge detection of automobile radar support assembly image based on Canny operator is determined,and the target contour extraction is realized.The realization of assembly quality inspection system of automobile radar support.Aiming at the three different assembly problems of automobile radar support,respectively,this paper proposes the principle of linear detection based on Hough transform to detect the missing assembly,the virtual welding detection based on the principle of Hough transform circle,and the assembly aperture detection based on least square fitting circle theory.In this way,the assembly quality of automobile radar support can be detected.In addition,based on MATLAB GUI software development platform and MATLAB Stereo Camera Calibrator,camera calibration and image processing algorithm programming and man-machine boundary are realized.Surface design.On the basis of theoretical research,the system has carried out a lot of experimental verification work.The results show that the system can complete the inspection contents of missing installation,virtual welding and assembly aperture size of automobile radar support.It meets the requirements of improving the efficiency of assembly quality detection of automobile radar support.