This Small Business Innovation Research (SBIR) Phase I Project proposes to investigate the problem of globally aligning multiple point scans of a 3D object. Given the digitized scan data for two partially overlapping surfaces, a numerical measure is described that quantifies the resistance encountered by one sliding over the other (lockability versus slippability). This measure is a heretofore missing piece in the design of successful algorithms for aligning multiple partial scans. It is of fundamental importance in building a prioritized schedule that optimally merges the most lockable scans. It is to be expected that a proper fine-tuning of this numerical method will dramatically improve the speed and accuracy of computed alignments, particularly for mechanical shapes. This yields a classification of the defining surface which enables the detection and fitting of primitive features in CAD models directly from the collected point data since these are generally slippable surface elements such as planes, spheres, cylinders, cones, surfaces of revolution and extrusions
