Vision Based 3D Map Building Indoor Environment

Abstract

In this paper, we propose a vision-based 3D map building method for indoor environments using monocular camera. The proposed method first extracts straight line information from images and calculates their equations in three-dimensional space. We use the Canny edge detector and the Douglas-Peucker Algorithm to extract straight line information. We also use Kanade-Lucas-Tomasi Feature Tracker (KLT) to obtain feature points at edges and corners, and Principal Component Analysis (PCA) method to find line equations. Second, it constructs plane information from line information.