On-Line Monitoring and Defects Detection of Robotic Arc Welding: A Review and Future Challenges

Robotic arc Welding is the main manufacturing technology for key structure components in the industries of aerospace, automobile, nuclear, ship and military equipment. Real-time monitoring, controlling and detecting of the welding process and seam quality can improve the stability and reliability of weld quality while increasing the efficiency and accuracy of defect detection. In this paper, we briefly reviewed the state-of-art on-line welding process monitoring based on different sensing techniques, including image vision, laser vision and distance, arc optical emission, arc audible sound and new immerging X-ray computed tomography. Then, a concise review of feature dimension reduction and selection is provided before the multisensory information fusion. The anticipated challenges are carefully discussed from the aspect of data correlation, evaluation and deep learning. We believe that more attention should be paid on topics such as real-time inner defects detection combining with defects micro characterization; problems related to complex-thin-big structure component welding; and applications of the latest deep learning technologies.