Environment Observation by Structure from Motion with an Omni-directional Camera

Measurement of a surrounding environment structure is important for mobile robots to make autonomous locomotion. In this paper, we propose a method for three-dimensional measurement of the environment using an omni-directional camera installed on a mobile robot. The method is based on structure from motion under the assumption that the environment is static. Along with the image sequence, it tracks feature points to get corresponding points for stereo matching. Using stereo pair images taken at selected observation points, the method estimates the relative relations of camera positions and orientations. With these relations and image coordinates of the feature points, three-dimensional coordinates of object points are calculated by triangulation. Integration of the individual measurement data realized by scale matching produces a whole map of the environment. Experimental results show the validity of the method.