Applying image processing technology for mobile robot automatically tracking and following steel weld seams

Abstract

For inspecting steel weld seams, most of the buildings, steel pillars, pipelines, storage tanks, etc. are currently still carried out in a manual way. More specifically, technicians often use scaffolding systems for moving and carrying defect testing equipment. At this time, the job requires them to work in dangerous altitude conditions, even in environments containing harmful gases. Therefore, this paper will be focused on developing a mobile robot in order to replace human in the work of carrying welding defect equipment. This helps to ensure worker safety, increase productivity and save costs. A computer vision is integrated into the robot to create a unified system to help it see, think. The robot uses computer vision to automatically identify steel weld seam and align the movement direction to follow the welding seam. In addition, with a special design using a permanent magnet wheel system, the robot has the ability to move flexibly on the surface of large steel structures with a flexible tilt angle from 0˚ to 180˚.