This paper presents a position estimation method for planetary rovers based on panoramic skyline images. The estimation is conducted by a comparison between two skyline images. One is extracted by image processing of an omni-directional image that is obtained by an omni-directional camera. The other is obtained by a geometry-based calculation from a digital elevation map that was obtained by satellites. To confirm a validity of our method, we applied the method to omni-directional images that was obtained in a desert. In this paper, we introduce our method, report the result, and discuss its resolution and availability.