Modelado 3D de terrenos a partir de fotografías

Abstract

There is increasing interest in three-dimensional modeling of land for various purposes. So far simulation techniques based mainly on graphics tools more or less complex than using terrain point clouds, simulated 3D mesh networks or adapted to the terrain. Other tools have recently appeared with the same objective, but whose philosophy is very different and that is to create three-dimensional models using these series of photographs with large overlapping areas obtained from different viewpoints. Using automatic detection algorithms and deduction homologous points of transformation parameters, is possible to deduce that land 3D points with relative accuracy. This gives the mesh that defines the digital terrain model and meshing that may overlap the image of the land and all the objects that are located on it. Recent tools like Autodesk 123D Catch, can fulfill this purpose perfectly. Our goal in this paper is to examine the application and analyze its suitability for generating 3D terrain model from photographs. We will try to compare the suitability of this program over other more traditional techniques. Obtaining land 3d models from photographs can be a good tool for the rapid acquisition of these models without having to resort to complex and costly procedures topographic. These models can be used to pre-planning studies, preliminary designs, presentations,…