Research On Robot Intelligent Welding Of Lap Joint Based On Laser Vision Sensor

With the rapid development of welding technology,how to ensure the welding quality of products as far as possible has become the concern of the majority of welders.At present,most robots work in the mode of teaching and playback,and their walking trajectory and technological parameters are fixed.Once the state of the workpiece changes,the robot cannot make corresponding adjustments,leading to the decline of welding quality and even the waste of the product welding.Aiming at the problem that the welding quality is decreased when the robot works in the mode of teaching and playback when the gap exists in the flat welding of lap weld,a laser vision sensing system is designed and built to improve the intelligent level of the robot,which has important application value for improving the welding quality of products.In order to improve the accuracy of lap joint detection,this paper adopts Zhang Zhengyou's plane calibration method to calibrate the camera,and obtains the internal and external parameters and distortion coefficient of the camera,thus correcting the distortion phenomenon generated when the camera collects images.The calibration re-projection error is less than 0.25 pix,and the accuracy can reach the sub-pixel level.At the same time,the“hand-eye” calibration of the robot was carried out,and the “hand-eye” transformation relationship of the robot was obtained.The algorithm flow of laser stripe feature point extraction for lap joint includes bilateral filtering,image thresholding,edge extraction,center curve extraction and feature point extraction.Aiming at the image thresholding processing,this paper proposes the method of determining different weights according to different positions of the image,which can reliably extract the laser stripes from the background and improve the accuracy of image processing.According to the projection principle of vision inspection,the welding seam with different gaps was reconstructed by 3D scanning,and the test was carried out by the algorithm designed.The results show that the error value of detection can be within 0.2mm,the average value is 0.104 mm,the maximum variance value of detection is 0.0541,and the image detection accuracy can reach 0.246 mm.After testing,the detection period of the system is less than 60 ms,and the detection speed is extremely fast.I/O communication and actual connection design are carried out between IPC and robot system,and the control software is modularized based on VC++6.0 platform.The rectifying test proves that the rectifying accuracy of the system in horizontal and vertical directions can be within 0.3mm and 0.4mm respectively,which reaches the advanced level in China.At the same time,a swing arc welding scheme based on laser vision “detect before welding” was designed in the intelligent control mode in this paper,which could adjust the welding torch height,the traveling speed and the swing amplitude according to the change of the lap joint gap.The results show that when the lap joint gap is between 0mm and 2.0mm,compared with the traditional teaching and playback mode,under the intelligent control mode,the joint area of upper and lower plates of lap weld can be improved effectively,and the surface and section of the weld are formed with good results.