Research on the welding strategy and welding technology for medium-thick plates based on three-line structured light vision

Weld reconstruction technology based on visual sensors is a crucial component of intelligent robotic welding systems. In thick-wall workpieces with big groove multi-layer automatic welding, precise measurement of the workpiece and weld seams directly affects subsequent weld path planning and posture planning of the welding torch. To address this issue, this study employs a three-line structured light measurement system for weld profile detection. A sub-plane fitting algorithm is proposed for 3D reconstruction of weld grooves, accompanied by a rapid workpiece coordinate system establishment method using three-line structured light to ensure accurate boundary segmentation. However, direct welding based on reconstructed feature points cannot guarantee high-quality results. Therefore, we investigated the effects of key process parameters, such as welding position, direction of welding, and torch inclination angle on weld bead geometry. A second-order general rotational composite design was implemented to develop regression models correlating welding parameters with weld bead characteristics. Finally, integrated process planning combining reconstruction data with optimized welding parameters was implemented for torch trajectory and orientation planning. Experimental results demonstrate that this methodology effectively satisfies quality control requirements for automated welding of thick-walled structural components.