Research On T-shapede Welding Automation System Based On Binocular Vision

As an important processing method in the manufacturing industry,welding has been widely used in industrial production.With the transformation of the manufacturing industry towards automation and intelligence in recent years,the demand for automation and intelligence in the welding process is also increasing.At present,welding robots are mostly teaching-playback robots,and their ability to adjust to external disturbances is poor,which seriously affects the quality of welding.In response to the above problems,this dissertation is based on binocular stereo vision technology,focusing on the welding seam detection technology in T-shaped welding,and building a T-shaped welding automation system.The main contents are as follows:(1)First,we introduced the basic principles of binocular stereo vision,studied the calibration method of binocular vision,calibrated the binocular camera and analyzed the source of error,used nonlinear methods to optimized the calibration results.(2)Then,the weld image is firstly processed by graying and ROI extraction using gray average to improve the signal-to-noise ratio,and we compared and analyzed the filter methods,used the median filter to denoise.On this basis,the dissertation proposed a method which combined the Hough Line transform and corner detection to position the welding spots,the position of the welding spot is obtained.(3)Secondly,the acquisition of the three-dimensional coordinates of the welding spot,and we proposed a method of firstly searching longitudinally and then matching horizontally to carry out the stereo matching,which improves the matching efficiency.After obtaining the 3-D coordinates of the welding spot,it is converted to the base coordinate system of the robot arm by hand-eye calibration,finally,we analyzed the error.(4)Finally,we designed and built a T-shaped welding automation experiment platform,wrote programs of each functional module of the software,and finally carried out welding simulation experiments on the experiment platform.By analyzing the error of the experimental platform results,the effectiveness of the T-type welding automation system proposed in this dissertation is verified.