Automatic seam detection and path planning in robotic welding

To make robotic welding systems more flexible, vision sensors are introduced as they provide large amount of information about the welding components. In this paper a method is introduced to automatically locate the weld seam between two objects in butt welding applications. The proposed method provides flexibility for robotic welding by having the ability to locate the weld seam on arbitrarily positioned work pieces. This method is also cost effective as it is developed using images captured from a low cost web-cam. Furthermore, the proposed method is able to plan a robot path along the identified seam. Simulation and experimental results show that the method can be used successfully in detecting and locating seams on variously shaped work pieces and robot paths can be successfully generated to follow the weld seams.