Review of Implementation of Vision Systems in Robotic Welding

Abstract

The changes in technology between Industry 3.0 to Industry 4.0 brought about a revolution of true automation. Many previously manual processes were extensively automated, and this allowed the development of specialized robots for purposes such as welding. While technology, in general, has advanced quite rapidly across the world, it is still unable to match some clear human advantages, like on-the-spot decision making and understanding the importance of context for a job. There is undoubtedly technology that successfully assists in manual welding, but there are no fully automated technologies of this type anywhere. This paper is a review of various research papers that have dealt with important concepts involved in advancing and improving robotic welding technology. It talks about the optimal software architectures for this type of job, and also the various methods of seam tracking for a weld joint. These range from real-time seam detection to a technique that compares weld seam images to a database and even identifies the type of weld joint. This is an important part of the process since it forms the basis of the input provided to the robot. Only a high-quality input will result in high-quality output. Additionally, this paper covers the various welding parameters that can be altered, such as voltage and feed rate.