3-D vision-based trajectory tracking of welding robots

Abstract

The trajectory accuracy of laser welding robots is improved by a robot trajectory tracking method using a robot control system based on visual feedback through two cameras mounted on the end of an industrial robot.A threshold segmentation method is used to find the laser focus and a gray projecting integral method is used to quickly detect weld seam.Then a modified line-line matching method and stereo vision technologies are used to calculate the laser focus and weld seam positions and the error between the laser focus and the weld seam.The robot then moves to reduce the error as soon as possible based on the robot kinematics.Test results show that the method effectively improves the trajectory accuracy of industrial robots.