Automatic alignment of multi-view range maps by optical stereo-tracking

Purpose: Industrial reverse engineering applications usually require reliable multi-view measurements of complex shapes obtained by full field techniques. One of the most challenging issues in this field is represented by the automatic computation of the best transformation parameters relating the different views into a common reference system. Method: A low cost optical tracking system has been developed with the aim at creating an automatic procedure to align 3D point clouds captured by a structured light system. The tracking system uses stereo images and retro-reflective infrared markers rigidly connected to the scanner. Markers are accurately tracked on the basis of automatic intensity-based analyses. Stereo correspondences are established by using epipolar and similarity constraints. Result: Experimental tests have been carried out in order to evaluate the capabilities of the developed system in the measurement process of nominal benchmarks and complex shapes. Discussion & Conclusion: The accuracy obtained in the automatic alignment of multiple views justifies the use of the proposed approach in the speed up of the measurement process for reverse engineering applications, especially when conventional high-end systems cannot be applied.