Vision processing and control of robotic Arc welding system

Abstract

A microprocessor-based control system is presented for a Gas Tungsten Arc Welding (GTAW) process to join thin sheet metal parts. The system uses a welding robot, a vision sensor, and an image processor to control the welding torch in real-time. A vision-processing algorithm is developed to compute weld puddle geometry parameters from the noisy image of the molten pool. The weld quality is controlled by regulation of puddle area, puddle width, arc length, and puddle center over the joint. Experimental results indicate successful operation of the control system with heat sink disturbances in the weld fixture.