Abstract

REVELO-LUNA, David Armando; USAMA, Francisco Daniel  and  FLOREZ-MARULANDA, Juan Fernando. 3D Reconstruction of Scenes by Means of a Stereoscopic Vision System, Based on Feature Extraction and Developed in OpenCV. Ing. Univ. [online]. 2012, vol.16, n.2, pp.485-500. ISSN 0123-2126.

Recently, devices that allow obtaining three-dimensional information from the scenes have been developed. The most popular is the one that have been created by a stereoscopic couple of cameras. This work presents a development in stereo artificial vision for getting 3D coordinates of a real scene, which is processed online. The stereo algorithms are designed using OpenCV and are supported by Linux. Disparity maps appear in gray scale and in pseudo color in order to represent the depth. The 3d information is represented by clouds of points in a graphical virtual environment. Measures of an object with known dimensions resulted in erros between 0.72% and 6.9%

Keywords : Machine vision; disparity; stereo vision; epipolar geometry.

        · abstract in Spanish | Portuguese     · text in Spanish     · Spanish ( pdf )