A Vision-based Robotic Offline Programming System For Deburring Process

In order to produce burr inevitably in the process of casting parts manufacturing,burr on the quality of parts,appearance and post-processing and many other aspects will have a bad impact.The traditional robot deburring is taught by hand or programmed off-line,but the casting parts occasionally have shape deviation,and the burr distribution and size of the parts are random,resulting in a high rate of defective products,or even the impact of the robot and the casting.Therefore,by studying image processing technology,this paper detects the position and size of workpiece edge and burr to provide data for trajectory generation and modification of robot off-line programming system,so as to improve the processing quality of robot burr removal and reduce the rate of defective productsFirstly,the whole deburring system and its principle are introduced systematically.In the deburr system,the robot adopts Japan's NACHI MC20-01 robot,path extraction and planning adopts the off-line programming system BW_OLP developed by the laboratory,and the camera of the vision module USES the CCD camera of image.Based on this,this paper builds a visual assistant robot off-line programming automatic deburring test system,which automatically adjusts the distance between the robot end and the burr according to the actual workpiece size to meet the requirements of polishing accuracySecondly,in order to accurately obtain the position and size of grinding edges and burrs of workpieces,after calibrating the camera,the machine vision system developed in this paper is used to analyse and process the collected images.After image pre-processing,several widely used image processing operators were compared,and image detection was carried out by using the best performance Canny operator for edge detection and HoughLinesP operator for straight line detection.By using the parameters of relative pose calibration,the pose of the image object is transformed into physical space,so as to calculate the edge position and burr height of the workpiece.Experimental results show that the measured data processed by this algorithm are consistent with the results of subjective evaluation,indicating that this visual system is effective and practical in image detection of robot deburring.Thirdly,according to the edge position coordinates obtained by edge detection,the off-line programming system automatically generates the initial trajectory of the robot.According to the detected burr height information,the processing track of burr removal is adjusted in real time to reduce the processing speed for larger burr,so as to completely remove burr.By analyzing the extracted burr height information,the mean and standard deviation of the burr height were calculated.The high speed range of the burr height in the mean+2x STD range was taken as the high speed range.The robot cut burr at a high speed.Beyond this range,the robot feeding speed is reduced to improve the burr removal rate.Finally,the robot motion program obtained after the off-line programming module adjustment is imported into the robot control cabinet.The experiment shows that the robot deburring processing quality can be effectively improved by visual extraction of edge information and off-line programming trajectory adjustment.