Analysis of remotely sensed pre- and post-disaster images for damage detection

This study proposes a method to utilize the satellite, aerial and other remotely sensed pre- and post-disaster imagery data to perform geometric modeling and correlational analysis on the reconstructed models in order to detect the change associated, for example, with major regional and/or individual structural damage. Correlational analysis often fails to detect structural damage when only input images are utilized, especially if images are acquired under different illumination conditions. In fact, automatic detection in such cases becomes extremely challenging since making distinction of change due to structural damage from that associated with the difference in the illumination condition is extremely difficult. Many researchers have tackled this difficulty and proposed some methods of solutions including recursive hypothesis testing procedure. Although these methods provide a very useful basis for change detection, their applications are not universality successful for a variety of reasons. In order to achieve the required level of accuracy for the proposed application and locate the site of detected damage, it is proposed to use available GIS maps to register remotely sensed images. It is further necessary that a user-assisted three-dimensional model be reconstructed and correlational analysis performed. The algorithm performs successfully for change detection. However, issue of occlusion remains as a challenge that requires further investigation.