Research On Visual On-line Detection System For Grinding Workpiece Defects And Roughness

With the rapid development of industrial robots and machine vision technology in the field of grinding and polishing of workpieces,and current of new era in which Chinese manufacturing industry is developing towards intelligence,automation and low cost.In this paper,study is commenced on the grinding automation technology and the machine vision detection technology related to the intelligent inspection of the work piece,and an efficient and practical solution is proposed.It is mainly divided into three aspects:automatic grinding of the workpiece,surface defect detection of the workpiece and grinding the surface roughness of the workpiece.In order to realize the automatic grinding and polishing of the workpiece and the real-time online intelligent detection of the workpiece,this paper combines the grinding industrial robot with the machine vision technology,moreover,on the basis of the automatic grinding of the workpiece by the grinding machine at the end of the robot arm that controls the grinding of the industrial robot,a visual on-line inspection system for grinding workpiece defects and roughness is designed.The research contents are as follows:(1)For the visual detection of ambient illumination unevenness,the collected images have local exposure,local shadow and other issues.A method is proposed by this paper for detecting the defects of polished workpieces under uneven light conditions,which firstly performs preprocessing such as graying and binarization on the acquired image to extract the main contour of the mobile phone shell;and then uses the localized binarization method to solve the problem of false detection of the defect of the mobile phone casing due to uneven illumination;and finally,the edge detection algorithm is used for defect detection to mark the defect location and size.(2)In the non-contact roughness measurement method,the light-cutting method has the advantages of high measurement resolution,simple measurement method,easy operation,low cost,etc.In this paper,the light-cutting microscope and the matching CCD camera are used to collect the roughness image of the polished workpiece;then,the workpiece roughness image is subjected to binarization,filtering and other pre-processing,and then the edge detection algorithm is used to obtain the roughness profile curve;the least square method is used to fit the curve to reduce the error,and the pixel height difference is obtained;finally,according to the correspondence between the image pixel height difference and the contour height value in the standard roughness sample,the surface roughness value is calculated.(3)Based on the research of grinding workpiece defect detection algorithm and roughness detection algorithm,this paper designs software for visual online detection system and conducts related experimental research.The system realizes the on-line detection of defects and roughness of the polished workpiece automatically,and displays the detection result and the detection effect map on the computer in real time.Experiments show that the defect detection part of the system can detect defects such as stains,scratches,scratches and marks,and the detection time is 1.6s/piece,and the accuracy rate is 95%.The roughness detection range in the system is 1um.~80um,the detection time is 0.7s/piece,which has good application value and economic value.