Vision-based sensing of the welding process: a survey

Abstract

A substantially successful technology that has emerged is the use of electro-optics, which has been expanded into the realm of 'machine vision' with the rapid advance in optical sensor technology and computer processing power. The basic principle of the optical sensor has remained unchanged over the years. Charge-Coupled Devices (CCDs), laser diodes and microprocessors form the basis of this type of sensing technique. CCDs have found their use in the observation of weld pool. It is known that the surface of the weld pool contains information that can be used to control the welding process. Researchers have used cameras to observe the weld pool and study the shape of the weld pool. Optical sensors have also been used to measure the arc light to determine the state of the weld pool. This paper summarises limited research done in the field of sensing the weld pool visually or optically and is intended to give the reader a brief overview of what optical technologies and techniques have been used to sense the weld pool state.

An Imaging and Measurement System for Robust Reconstruction of Weld Pool During Arc Welding

“…Fig. 11 shows the result from [27], where only some lines are shown. These lines did not reflect the depth of the weld pool correctly and enlarged it significantly.…”

“…13 shows the reconstructed results from [38], and only some points are reconstructed. Its resolution is also too low to give a meaningful weld pool surface as [27], although its solution is analytical and accurate.…”

Binocular vision system for both weld pool and root gap in robot welding process

“…Thus, visual sensing systems have been developed to view the weld pool. Vision sensors applications in welding process have drawn significant attention in welding research community recently (Saeed, 2006), because vision sensors can provide abundant information, including the joint type, initial welding position, real‐time states of weld poor and root gap, and so on. Chen et al (2006) use vision sensor to detect the initial weld point and the joint type, and thus provided weld seam information for guidance.…”