Plural pieces of frame data are generated and held based on image data obtained by capturing images while moving a plurality of capture devices for capturing images in different directions. A frame data group containing consecutive pieces of frame data determined to have been captured at practically the same position can be extracted from the plural pieces of held frame data. Frame data is deleted from the extracted frame data such that one piece of frame data can exist as the data captured at one position. Thus, when the frame data corresponding to a plurality of positions on a map is stored, thinned frame data is stored after being associated with the corresponding position on the map. Therefore, when virtual space in a walk-through system is designed using an IBR technology, panoramic images for reproduction of the virtual space can be efficiently generated and held.