A detection and configuration method for welding completeness in the automotive body-in-white panel based on digital twin

Abstract

To address the problems of poor welding completeness and inefficient configuration for defective automotive body-in-white panels, we propose a method for detecting and configuring the welding completeness of automotive body-in-white panels based on digital twin (DT) and mixed reality (MR). The method uses DT to build an MR-oriented DT framework for the detections and configuration of body-in-white panel welding completeness. We propose a method to build a DT knowledge base for panels, a Yolov4-based welding completeness detection method, and a MR-based configuration method for the welding completeness in panels. Our team develop a panel welding completeness detection and configuration system to fully validate the effectiveness of the method.