Vision sensing for arc welding robots-a new approach

Abstract

A brief review of the sensor systems used in seam tracking for arc welding robots is represented. The authors present a new approach to the problem which makes use of the welding arc as the light source to illuminate the weld seam. Single-line images perpendicular to the direction of movement of the arc gun are analysed for grey-level distribution. Various forms of data processing are assessed as a means of extracting the seam position. Averaging, cross correlation and matched filtering are shown to be equivalent for the special case of butt welds. It is shown that the seam position can be extracted from groups of single-line images using these different filtering techniques.