Ultrashort pulsed laser beams are used for numerous surface processes, such as surface functionalization, thin film processing, or micro texturing. Laser beams in combination with three-axes galvanometer scanners offer the potential to address complex (free form) surfaces. However, the complexity of machine calibration, path planning, and process controller increases significantly compared to flat field processing. In this work, we present a triangulation-based surface recognition method to localize the workpiece within the scanner coordinates. The surface data is fed into a real-time process controller which performs accurate coordinate transformations from model space to the actual workpiece surface. Performing coordinate transformations based on surface measurements reduces static positioning errors of the laser beam on the surface. Furthermore, by computing coordinate transformations in real-time enables integration of additional sensor data. In particular, incorporating position encoder signals from the stage system may lead to synchronized laser scanning processes of complex surfaces on moving objects.