3D reconstruction and measurement of indoor object using stereo camera

Abstract

This paper implements a study of 3D reconstruction and measurement of indoor objects using binocular stereo camera. First, SIFT feature are extracted from both images and the matching points are found. RANSAC method is applied to eliminate wrong matching points. Second, two stereo images are rectified and generated disparity mapping. Finally, the depth mapping and 3D information of each pixel are derived. This method is applied to reconstruct and measure the true object in indoor environment.