Resumo

GRANDON-PASTEN, Natividad; ARACENA-PIZARRO, Diego  e  TOZZI, Clésio Luis. 3D OBJECT RECONSTRUCTION WITH CALIBRATED IMAGES. Ingeniare. Rev. chil. ing. [online]. 2007, vol.15, n.2, pp. 158-168. ISSN 0718-3305.  doi: 10.4067/S0718-33052007000200006.

The system is composed of two main modules. The first one, carries out the image prosecution, whose objective is to determine the depth map of a pair of views where each pair of successive views continues a sequence of phases: interest points detection, points correspondence and points reconstruction; in the reconstruction process, is determined the parameters that describe the movement (rotation matrix R and the translation vector T) between the two views. This an sequence of steps is repeated for all the peers of successive views of the set. The second module has as objective to create the 3D model of the object, for it should determine the total map of all the 3D points generated, by each iteration of the previous module, once obtained the map of total depth generates the 3D netting, applying the triangulation method of Delaunay [28]. The results obtained of the reconstruction process are shaped in a VRML virtual environment, to obtain a viewing more realist of the object.

Palavras-chave : Computer vision; 3D reconstruction; matching; union of views; 3D visualization.