Research On Weld Seam Recognition And Welding Path Planning Of Welding Robot System

Structured-light vision measurement technology is widely used in the field of welding robots because of its high measurement accuracy,high speed,non-contact,and low cost.The new generation of welding robots trend to replaces traditional manual welding and teaching-reproducing welding robots.The accurate calibration of the structured light vision measurement system,the accurate recognition and positioning of the welds,and the reasonable planning of the welding trajectory are the core issues of this kind of welding robot system.In this paper,the six-degree-of-freedom robotic welding system,the calibration technology of the visual measurement of the structural light,the image recognition technology of the weld image of the strong reflective material,and the welding trajectory planning method of the fillet weld are studied in depth.The main work of the paper are as follows:(1)The stereoscopic measurement principle of structured light vision measurement technology is studied,its measurement model are divided into three parts: camera model,structured light model,and hand-eye model.The calibration method of internal and external parameters of camera and the calibration method of structural light plane parameters are studied respectively.The calibration method and optimization method of hand-eye matrix are deduced.(2)Detection of feature points of the weld image.Aiming at the strong reflection characteristics of the stainless steel fillet weld image,by analyzing the image of the weld taken under different robot attitudes,a method for detecting the weld feature point based on image morphology was proposed.The image connectivity filter and image ROI(Region of Interest)extraction,corner detection and other methods are used to extraction of the weld feature points.Secondly,the weld 3D reconstruction and robot welding trajectory planning methods are studied.Based on the structural light vision measurement principle,the mathematical transformation of the pixel coordinate system to the robot coordinate system is deduced,which provides a theoretical basis for the 3D reconstruction of the weld seam.Thenon-uniform rational B-spline interpolation method is studied.The point cloud data was spline-interpolated,and the transformation from discrete trajectory to continuous trajectory is realized.According to the specific weld shape,the planning algorithm of the welding torch pose is studied.(3)By using the open source software libraries such as Qt,OpenCV,and PCL(Point Cloud Library)and multi-threaded programming,the six-degree-of-freedom robot welding system software based on the Windows platform is designed,and the coordinated control of robot system,integrated control system,welder system,and structured light vision measurement system are implemented.The designed software has functions such as human-computer interaction interface,algorithm running,and data recording.(4)Weld seam tracking experiment and welding experiment,designed and built an experimental platform,and verified the accuracy of the structured light vision measurement system by measuring the target data at a fixed position.By measuring and tracking the actual welds,the weld identification method,welding trajectory planning method,and feasibility of robotic welding system are verified.Finally,actual welding tests are performed on a batch of stainless steel samples.The experimental results show that the designed structured light vision measurement system can realize the measurement and tracking of the stainless steel fillet welds.The maximum error of the measurement space point relative to the robot coordinate system in the X-axis direction is 0.73 mm,and the maximum error of the Y-axis is 0.65 mm and the maximum error of Z axis is 0.59 mm.This measurement accuracy meets the automatic welding needs of the actual fillet weld.