複数の全方位画像上の床領域を統合した移動ロボットの自己位置推定

抄録

A self-localization scheme for mobile robots based on model-fitting of a floor region captured by omni-directional vision is proposed. Even if parts of the floor region hidden by obstacles and walls, these can be compensated for by computing improved convex hull of a boundary point set of the floor region and by integrating multiple floor regions on the omni-directional images.