Improving position accuracy in large-scale laser structuring processes using surface feature detection algorithms

Laser structuring is an effective way to functionalize or texture components. For the machining of large components > 500 x 500 mm, precision machine tools costing up to €1.5 mio are needed. This circumstance leads to a limited market applicability. Fraunhofer IPT is developing a cost-efficient laser structuring solution making it possible to scale up maximum processing area by mounting the laser structuring head on an industrial robot. However, high positioning accuracy of < 5 µm is needed for laser structuring while an industrial robot system delivers a maximum position repeatability of about 70 µm. Fraunhofer IPT has developed a compensation strategy using an inline camera setup. The system detects reference features of markers and texture by using FAST algorithm for corner detection with a laplacian image pyramid. The current system is capable to set exposure times automatically and to compensate position inaccuracies of up to 1 mm to an overall accuracy of < 5 µm despite a camera chip resolution of 5.3 µm/px. In this contribution we present an overview of the compensation algorithm with an evaluation method for process stabilization and identify challenges, e.g. illumination and material properties will be addressed.