A Method for Registration of Urban Aerial Images Taken before and after a Disaster

Many methods of automatic analysis of wide urban areas after a large-scale disaster using aerial images have been researched. In several of these methods, information is obtained by integrating aerial images taken before and after the disaster or by integrating the images and a map. For their automation, images must be registered automatically. In this paper, we propose a new method of automatic registration of urban aerial images taken before and after a disaster. First, this method detects feature points whose changes are few in the images taken before and after the disaster. Then, this method registers the images automatically by using the locations of the feature points and a_geometric transformation.