Abstract
In the Tohoku Earthquake Disaster 2011, the tremendous efforts were required to understand the liquification damage in the vast area. For the future, the automated and standardized approach to detect the liquification damage is expected. In this study, several approaches to detect the sand boil caused by liquification were examined. The high resolution satellite images such as World-View-2 which were taken before and after the disaster were used. The image was classified based on both the pixels and the objects, using Clustering, Decision-Tree and Maximum Likelihood methods. The result showed that Maximum Likelihood method based on the object oriented image was most effective to detect the sand boil.[This abstract is not included in the PDF]
