The laser welding process for automotive body structures demands a high level of seam recognition and localization. Swift and accurate identification and positioning of seams are essential technical means to ensure an intelligent welding process. Seam tracking during the automotive body laser welding process relies on real-time sensing by sensors on the workpiece, determining seam positions, and feeding this information back to the welding robot’s motion control system to adjust robot posture and motion path promptly. Therefore, the precision and real-time performance of seam tracking are crucial for ensuring welding quality.
Seam tracking technology primarily encompasses two phases: pre-weld seam detection and real-time in-weld seam tracking. Both phases rely on visual sensors to identify and locate seams. Pre-weld seam detection aims to provide initial welding motion paths to the welding robot. In-weld seam tracking’s objective is to accurately identify seam positions and dynamically adjust the robot’s motion trajectory. Currently, seam tracking technology is applicable to various laser welding processes for automotive body structures, displaying a certain degree of generality.
At present, based on visual sensing technology, especially laser visual sensing technology, seam tracking methods have gained extensive research and application due to their advantages of high contrast, high precision, real-time performance, and non-contact nature.
The accurate detection and recognition of pre-weld seams in the automotive body laser welding process play a crucial role in guiding the trajectory planning of welding machines. The precision of seam detection directly affects the accuracy of welding robot paths. Addressing the characteristics of thin-sheet welding in automotive body structures, Wang Bangguo employed structured light computational methods to plan welding trajectories. By utilizing visual sensing technology, technical support was provided for tracking pre-weld seam information. Structured light-based visual sensing technology achieves precise seam recognition, and when combined with seam recognition algorithms, it enables accurate identification of pre-weld seams.
Real-time detection of seams in the laser welding process for automotive body structures enables dynamic adjustments of the welding robot’s motion trajectory. This helps minimize seam positioning errors and enhances welding quality. HGTECH applied a three-dimensional laser vision system to the precision control of roof laser welding seams in the automotive white body assembly process. The system is an advanced setup centered around computer, information processing, image processing, and laser vision technology. It demonstrated success with JAC Motors resulting in the M111 project’s white body roof welding production line application. The system achieved pre-weld seam tracking and pre-processing, compensating for welding errors by adjusting welding robot trajectories. This significantly improved welding quality and production efficiency.
The seam tracking technology in the automotive body laser welding process realizes pre-weld initialization of welding robot trajectories and dynamic adjustments during welding, leading to a more stable and reliable laser welding process, thereby ensuring welding precision.