Laser vision system for automatic seam tracking of stainless steel pipe welding machine (ICCAS 2007)

This paper describes an automatic weld seam tracking method of stainless steel pipe welding machine. Due to the complexity of the pipe groove shape, it is difficult to find a seam and operate a torch manually in the pipe welding process. So, the laser vision system for seam tracking is designed for sensing the seam position and controlling a torch automatically. To achieve precise seam tracking, the design of sensor head, image processing, and calibration are carried out. The measurement principal of a vision sensor in this system is a distance measurement using an optical triangulation technology. A structured laser light is projected on an object. A CCD Camera captures the reflected laser stripe and a position change of a laser stripe in the image, which is related to the 3D position of an observed object, is found. By performing camera calibration from 2D image coordinates and 3D object coordinates, position information of the object is acquired. The pipe welding machine is composed of a laser vision system, a welding controller, and a system controller. A laser vision system consists of a sensor head and controller. A sensor head is attached to the welding machine and acquires seam image containing a laser stripe. A controller receives the image and extracts a seam point. Finally it is transferred to the system controller through a RS-232 communication. The experiment result shows that this vision sensor works effectively. It will provide more precise welding performance and convenience to the operator.