Research On Automatic Glue Coating Intelligent Detection System Based On Machine Vision

Modern automobile manufacturing requires higher and higher testing requirements for gluing quality.Due to low efficiency and inconsistent error judgment standards of manual measurement,this direct contact measurement method can no longer meet the online testing requirements of modern automobile industry.With the rapid development of machine vision,the measurement technology of machine vision has gradually been studied and paid attention to in the field of industrial automatic detection due to its fast detection speed and high measurement accuracy.In order to improve the on-line real-time inspection of the coating quality of the vehicle body,this paper designs an automatic coating intelligent inspection system based on machine vision,and develops a vision measurement system taking the coating track of the vehicle side wall as the measurement object.Firstly,according to the measurement index of the gluing track size,each workstation in the gluing system architecture is designed,and a side wall gluing detection system centered on the machine vision system is built.The hardware selection of the vision measurement system is completed,the lighting scheme of the vision measurement system is designed,and the vision online measurement system platform is constructed.Secondly,based on the research of image preprocessing technology and traditional pixel-level edge detection algorithm,an improved C anny algorithm is used to extract the rough location of the edge features of the gluing track,and experiments show that the improved gluing edge algorithm is obviously superior to the traditional edge algorithm.Using morphological skeleton thinning extraction algorithm on the coarse positioning edge,the edge of single pixel gluing track is obtained.Based on this,the sub-pixel edge detection algorithm of cubic curve fitting is continued to realize sub-pixel fine positioning of gluing track edge.Then,the camera calibration technology of the measurement system is studied,the principle and implementation process of Zhang Zhengyou calibration method are clarified,the monocular camera calibration is completed and the camera calibration results are obtained.The physical size of the gluing track is calculated by the standard part calibration method,and the gluing track size measurement of sub-pixels is realized.Finally,the software design of the vision measurement system is completed,and the man-machine interface of the machine vision online measurement system is developed.The measurement accuracy of the side wall gluing system is verified through experiments.The experimental results show that the gluing system designed in this paper has higher measurement accuracy and faster processing speed,and can meet the real-time online detection requirements of the automobile industry on the gluing quality performance.