Weld seam profile extraction of T-joints based on orientation saliency for path planning and seam tracking

In the development of intelligentized robotic welding, weld seam profile extraction is a prerequisite in thick plate welding for path planning and seam tracking. This paper presents a robotic welding system to implement autonomous welding for T-joints, where a novel vision sensor is employed to capture weld seam profiles and weld pools simultaneously in the same frame. It exploits lasers to profile the T-joint. This paper concentrates on accurately extracting the seam profile from the background of the weld pool. An effective procedure of extracting the weld seam profile of the T-joint based on the orientation saliency of the laser stripe is proposed. Experimental results showed the effectiveness of the proposed method.