Research On Location And Fastened State Recognition Of Ring Groove Rivets Based On Machine Vision

At present,the riveting quality of ring groove rivets is judged by analyzing the riveting pressure-displacement curve.This method is easily affected by external factors such as friction and temperature,and cannot accurately identify the fastening state of ring groove rivets.Combined with the existing problems in riveting quality inspection,according to the principle of machine vision positioning technology and detection technology,this paper studies the ring groove rivet positioning and fastening state identification technology,and realizes rivet positioning and fastening state identification through the following steps.Firstly,the characteristics of the target workpiece are analyzed,the realization method of ring groove rivet quality inspection is determined,and the selection and installation method of hardware equipment in the machine vision system are studied.Secondly,the imaging model of the camera in the hardware equipment and the causes of imaging distortion are analyzed,and the Halcon software is used for camera calibration and distortion image correction.Then,a series of image preprocessing algorithms are used to process the positioning images and fastening state recognition images to obtain a clear target connected domain of the binary image.The target connection domain in binary image is segmented and identified by image feature screening,and the target connection domainof the positioning images and fastening state recognition images is located and identified respectively.Finally,the software platform of the ring groove rivet positioning and fastening state identification system is developed,and the experimental operation is carried out by using the established software platform,and the experimental results are analyzed.In order to improve the positioning accuracy and state recognition accuracy of ring groove rivets,the following improvements are proposed in this paper.(1)According to the characteristics of the research target,a positioning algorithm based on correction coefficient is proposed for the ring groove rivet positioning part.The experimental results show that the improved positioning method can improve the positioning accuracy of rivet coordinates.(2)For the ring groove rivet status identification part,a denoising method based on ROI of the region of interest is proposed.It is verified that the method can effectively filter out noise and improve the recognition accuracy.The results show that the rivet positioning accuracy rate of the machine vision-based ring groove rivet positioning and fastening state identification system is 97%,and the rivet fastening state identification accuracy rate of this system is 96%,which prove that the system can meet the riveting quality inspection requirements of the actual production line.