全方位画像中の特徴点と直線エッジの同時拘束と基線長の自動決定に基づくStructure from Motionによる三次元環境モデル生成

概要

論文の詳細を見る
In this paper, we propose a method for 3-D environment model construction based on structure from motion by using an omnidirectional camera. Map information is important for path planning and self-localization when mobile robots execute autonomous tasks. In an unknown environment, mobile robots should measure the environment and construct its map by themselves. Our proposed method uses point and line features to measure environments densely. Point and line feature-combined constraint condition make it possible to estimate camera movement precisely. Moreover, the proposed method optimizes baseline length for precision and robustness of camera movement estimation. The baseline optimization is invariant for environments, the number of features and camera movement. Experimental results show the effectiveness of our proposed method.