Online Welding Deviation Detection and Burn-Through Identification of Yolov5 Sheet Lap Mig Welding Based on Passive Vision

Abstract

Passive vision technology for quality condition inspection of metal inert-gas welding has been a hot topic of research in industry and academia. However, it is very easy to induce a series of defects during the lap welding of the sheet when using MIG welding. To overcome this limitation, this paper proposes a weld deviation detection and burn-through identification method. Firstly, a MIG welding molten pool image vision system is designed. Then, a new method for calculating the weld deviation is proposed, which is achieved by comparing the difference between the center of the lap position of the sheet and the center of the molten pool during the welding process. Finally, a comparison experiment was conducted to verify the reliability of the model for the identification of welding deviation and burn-through. This paper can provide some guidance for the online inspection of the welding manufacturing process.

Keywords: MIG welding, Yolov5；Passive vision technology；Welding defect detection