3D vision for automated structured welding

Robotic welding is widely used in modern manufacturing because it improves productivity, repeatability, weld consistency, and operator safety. Industrial robots are especially valuable in welding tasks because they can perform repetitive motions with high precision and reduce human exposure to heat, fumes, radiation, and hazardous working conditions.



However, many robotic welding systems still rely on predefined welding paths created through manual teaching, CAD models, or offline programming. These methods assume that the workpiece position and weld-joint geometry remain consistent between programming and production.

ABB Robotics Vietnam provides the industrial context for this project, including the ABB IRB 1300 robotic welding platform, RobotStudio environment, and the requirement to explore 3D vision-assisted weld-path generation. The proposed system uses a Revopoint 3D camera to capture point cloud data of the weld joint, while ABB RobotStudio provides the environment for path validation, reachability checks, collision detection, and RAPID program generation. The purpose is to connect 3D seam recognition research with an implementable ABB scan-to-path workflow.