For the autonomy of a mobile robot it is needed to know the position and orientation of a robot. Various methods of estimating the position of a robot have been developed. However, it is still difficult to localize the robot without any initial information. In this paper, we present how to make the colored map and calculate the position and direction of a robot comparing the map with an omnidirectional image. The wall of the map is rendered with the corresponding color images and the color histograms of images and the coordinates of feature points are stored in the map. Then a mobile robot gets the color omnidirectional image at random position and orientation, segments it and recognizes the feature points using the color histograms and the spatial arrangement of colorful regions. Among the recognized multiple fearture points the most reliable features anr selected and the position and orientation of the robot is calculated using the angle information of the selected features.
Color information is indispensable considering the geometrical symmetry of indoor corridor. The color information of objects can be effectively represented
using color histograms and the objects can be recognized using histogram matching. Based on the object recognition, the robot cna get reliable feature points. Proposed algorithm is very flexible to incidental case because of not using the initial position and orientation of a robot. And it is robust to rotation of the robot by using the omnidirectional sensor and color information.
Also the complexity of the map is not high because the map is composed of the color histograms and coordinates of the color objects.