EMPLEO DE IMÁGENES DIGITALES PARA RECONSTRUCCIÓN TRIDIMENSIONAL DE COMPONENTES MECÁNICOS

Abstract

The obtaining of a three-dimensional reconstruction from two-dimensional images presents a significant problem in several scientific fields. However, the high cost of the actual laser scanner systems leads to search other alternatives that permit to obtain a three-dimensional reconstruction with certain accuracy and economic cost. This work presents three alternatives for the three-dimensional reconstruction of the surface of different objects. Based on the fundamentals of the stereoscopic vision, the three-dimensional representation of an object is obtained from several photographs taken from different positions. For this purpose, the reconstruction methods using the software applications Patch Multi-View Stereo (PMVS), Microsoft Photosynth and Autodesk 123D Catch are analyzed and compared. These methods permit to take into consideration the geometric and photometric information of the analyzed object, which permits to obtain a representation of the characteristic parameters of the object very close to the visual perception of the object in the reality. Keywords: Photometry; geometric modelling; PMVS; Photosynth;123D Catch