Localization of mobile robots from full detection of ceiling outlines

In this paper, we propose a new localization system using ceiling outlines. We acquire the entire ceiling image by using a fisheye lens camera, and extract the lines by binarization and segmentation. The optical flow algorithm is then applied to identify the ceiling region from the segmented regions. Finally, we obtain the position and orientation of a robot via the center position and the momentum of the ceiling region. Since we use fully detected outlines, the accuracy and reliability of the localization system is improved. The experimental results show the effectiveness of the proposed method.