Research and prospect of welding monitoring technology based on machine vision

Welding monitoring technology based on machine vision has been widely researched in academic and industry, especially in the background of Industry 4.0, in that it can contribute to welding quality and productivity improvement. This paper outlines the technical points of welding status monitoring based on machine vision, including hardware and software. First of all, in the hardware part, the active and passive vision systems are briefly introduced, as well as the key steps in experimental deployment, such as the configuration of optical sensors and optical filters based on different detection objects. Secondly, some related image processing techniques in welding monitoring are also comprehensively reviewed. Additionally, the observed objects and their morphological characteristics of vision-based welding process monitoring are enumerated. On this basis, a series of intelligent models as well as optimization methods for recognition and classification in visual monitoring are considered in detail. Finally, potential research challenges and open research issues of welding visual monitoring are discussed to present an insight into future research opportunities. The main purpose of this paper is to provide a reference source for the researchers involved in intelligent robot welding.