3D Face measurement and scanning using digital close range photogrammetry: evaluation of different solutions and experimental approaches

This experimental research work was aimed to develop a 3D photographic method to be used in orthodontics diagnoses, having particular characteristics: precision of measurement, robustness of the method, speed of relief, ease of scanning, portability and low weight of the final system, low cost, non-invasive equipment. Various solutions of the experimental set-up have been implemented for the non-contact detection and measurement of the face soft tissues. This paper highlights the advantages and disadvantages of different set-up designed during the trials, comparing them to identify the best solution that meets the characteristics listed above. Mainly two different types of photogrammetric approach were designed: the first one uses three cameras and projected grids in a dark environment; it is based on point triangulation. The second one uses 4 or 5 cameras and operates with applied on the background and on the face in ambient light; it is based also on the DSM methodology, and obtains more dense point clouds. The systems were tested scanning the head of a mannequin and real human faces. Within each type, several variants were tested to evaluate the differences, changing the conditions of projection, lighting, shooting. Moreover, the results for each approach were compared with those obtained by laser scanning (both on real face and on a mannequin), for which are already known the performances. The system that best meets the requirements for diagnostic use in orthodontics proved to be the second one, in which coded targets applied directly on the face were used, shooting in ambient light, adopting digital photo sensors having 10 megapixels, and using 4-5 shots synchronized. The commercial software PhotoModeler was used for the CAD reconstruction of the face, and Geomagic for measures and comparisons.