Ripple Line Fillet Seam Tracking with Visual Sensors in Welding Process

Abstract

To overcome the problems resulting from weldment trimming and assemble errors, thermal deformation, arc blow, etc., a set of new tracking system of fillet seam for automatic corrugated plate welding was put forward based on machine vision technology. The tracking system included two cameras: the precursor was set in front of the welding torch (a corrugation cycle far away), for capturing the linear laser projected on the side plate and the base plate; the another one was installed over the melting pool for capturing the image of welding torch tip and the seam. The processed image information from the precursor camera was used to track the fillet seam three-dimensional trajectory and the seam width, according to the information the welding motion controller and the welding power made preplanning; the processed torch tip and seam image information was sent to the motion controller and the welding power so that they could make real-time motion and welding parameter adjustment. Computer simulation proved that this new tracking system could work efficiently.