Machine vision aided design for remote laser processing

Galvanometric scanners have become a frequently chosen industrial solution for remote laser processing of a wide range of materials. The scanners bring great advantage over conventional kinematics by delivering outstanding dynamics and processing speeds. That enables utilization of high power lasers and production volume increase by minimization of processing times. However, in batch production, as well as in small series or R&D tedious workpiece positioning and use of fixtures draws the limits on time saving. Here, we introduce developed modules for CAD/CAM systems, tailored to support remote laser scanning systems by enabling process design ""directly on the sample"". The presented solution uses orthogonal workspace imaging and raster vectorization for aided geometry fitting and adjusting process trajectories to the real position of workpiece. The developed algorithms bring flexibility and integrate designing stage with process realization. The system can adapt execution task to misaligned workpiece and still keep defined precision. This minimizes operation times and process failures. System errors were estimated experimentally on geometrical primitives. As a system validation, two industrial use-cases are presented, where material surplus removal is required. Overall system performance meets these tasks' accuracy requirements. The work focuses on research environment and gives basis for industrial architecture as further development.