Abstract
Three dimensional visual information is very important to recognize the location such as underground shopping area for the pedestrians, therefore 3-D map is effective for pedestrian navigation system instead of 2-D map. This paper proposes the system which creates 3-D map while walking. The 3-D information is calculated from omnidirectional images captured by a Hyper Omni-Vision, and the 3-D map is generated from the calculated 3-D information. As the fundamental experiment, the depth image is created from two omnidirectional images for creating the 3-D map.
