Obtenção de estrutura tridimensional a partir de movimento de câmara

Abstract

The acquisition of the three dimensional structure of a scene plays a major role in computer vision, where the knowledge of the 3D co-ordinates of specific scene components is often required. Examples of such instances occur in autonomously guided vehicles and robots, in the tracking and identification of moving objects, in automatic visual inspection, and so on. The general approach adopted in this thesis for capturing the three dimensional structure of a scene from a monocular image sequence includes the camera calibration, the detection of the selected entities in each frame of the sequence, the tracking of those entities along the sequence, and, finally, the computation of the 3D co-ordinates of specific points in the entities and the estimation of the depth co-ordinate of the other points.