BUILDING DAMAGE DETECTION OF THE KUMAMOTO EARTHQUAKE UTILIZING IMAGE ANALYZING METHODS WITH AERIAL PHOTOGRAPHS

Abstract

 It is important to detect damages immediately after the earthquake in terms of efficient disaster responses. In this point, we developed image analyzing methods by utilizing aerial photos of the foreshock and the aftershock of the 2016 Kumamoto earthquake. The image analyzing method consists of the differential analysis with digital surface models (DSM), the texture analysis, and the extraction method of blue tarps based on the region segmentation. With these methods, we extracted more than 15,000 damages of buildings caused by the Kumamoto earthquake, and compared with the visual photographic interpretation result which classified buildings with 4 damage levels. Then, we clarified that the DSM and the texture analysis are superior to detect destroyed buildings, and the extraction method of blue tarps is effective on the partially extraction of intermediate damages.