Applications of Molten Pool Visual Sensing and Machine Learning in Welding Quality Monitoring

Abstract

On-line monitoring of welding quality is very important to the realization of intelligent welding technology and has become a research hotspot in the field of welding technology. This article reviews the research results and latest research progress of welding quality on-line monitoring based on molten pool visual sensing in recent years. First, it introduces the characterization of welding quality by the two-dimensional geometric features and three-dimensional topographic features of the molten pool in detail, and then analyzes the application of machine learning and feature engineering in the online prediction of welding status; we discuss the deep neural network and welding quality at the end Convergence of online detection technology. The work done in this paper reviews the progress of online monitoring technology for welding quality and provides a basis for the follow-up work.