Improvement of Metric Accuracy of Digital 3D Models through Digital Photogrammetry. A Case Study: Donatello's Maddalena

In spite of an undoubted richness of information produced by 3D optical technologies, in some cases, the method for generating a digital model from single 3D acquisitions involves the propagation of errors. These errors limit the overall metric accuracy attainable with such procedure. This happens when small 3D images are assembled together in order to model a large object. The authors present a procedure by which the metric reliability of the 3D model, obtained through iterative alignments of single range maps, can be guaranteed to an acceptable level. For this purpose, non-impeding optical targets were specifically designed for placement around the object. These are measured using a close range digital photogrammetry technique and by the 3D range camera system. From these measurements, transformation matrices have been calculated. Each matrix allows for the roto-translation (pose) of the 3D images from the local coordinate system of the range camera to an accurate global coordinate system determined by the digital photogrammetric procedure.Malgr\ue9 la richesse certaine des informations produites au moyen des techniques optiques 3D, dans certains cas, la m\ue9thode de g\ue9n\ue9ration d'un mod\ue8le num\ue9rique \ue0 partir d'images 3D isol\ue9es entra\ueene la propagation d'erreurs. Ces erreurs restreignent la pr\ue9cision m\ue9trique globale qu'il est possible d'obtenir avec de telles m\ue9thodes. Cela se produit en particulier lorsque de petites images 3D sont assembl\ue9es ensemble afin de mod\ue9liser un gros objet. Les auteurs pr\ue9sentent une m\ue9thode gr\ue2ce \ue0 laquelle la fiabilit\ue9 m\ue9trique du mod\ue8le 3D, obtenu par alignement it\ue9ratif d'images \ue0 \ue9chelle unique, peut \ueatre garantie, ce \ue0 un niveau acceptable. \uc0 cette fin, des cibles optiques non masquantes ont \ue9t\ue9 sp\ue9cifiquement con\ue7ues pour \ueatre plac\ue9es autour de l'objet. Ces cibles sont mesur\ue9es au moyen d'une technique de photogramm\ue9trie num\ue9rique \ue0 distance rapproch\ue9e, ainsi que gr\ue2ce \ue0 une cam\ue9ra de t\ue9l\ue9m\ue9trie 3D. \uc0 partir de ces mesures, des matrices de transformation ont \ue9t\ue9 calcul\ue9es. Chaque matrice permet la translation-rotation (pose) des images 3D, depuis le syst\ue8me de coordonn\ue9es local de la cam\ue9ra de t\ue9l\ue9m\ue9trie, vers un syst\ue8me de coordonn\ue9es global pr\ue9cis d\ue9termin\ue9 par la m\ue9thode de photogramm\ue9trie num\ue9rique.NRC publication: Ye