INTELLIGENT 3D SURFACE DIGITIZING APPLYING OPTICAL RANGE SENSORS

Tri-dimensional surface digitizing applying optical range sensors is an important task in Reverse Engineering and Inspection processes. Such task normally requires a human operator intervention to define the appropriate viewpoints as well as the respective digitizing trajectories. As a consequence, final results depend on operator expertise and not on quantitative and objective methods. In this article a method to objectively select the best viewpoints and define collision free digitizing trajectories is proposed. The method is based on a coarse volumetric model implemented through a 3D voxel map. During digitizing process the model is properly updated in order to know which surface regions have been digitized and the quality of the acquired points. Viewpoint selection is performed based on the current model status according to a certain objective function, and the digitizing trajectories can be automatically defined to avoid potential collisions. Experimental tests performed in a typical industrial set-up system are reported.