Single Post-Event PolSAR Data Based Earthquake/Tsunami Damage Information Extraction in Urban Areas

An unsupervised earthquake/tsunami damage information extraction method using single post-event polarimetric SAR data is proposed in this paper. All urban areas are extracted by an improved classification method which can effectively discriminate urban and foreshortening mountain areas, they are grouped into four kinds of buildings to detect damaged buildings. The proposed unsupervised method can explicitly detect all the damaged buildings, including those with large orientation angle which are easily recognized as undamaged buildings. The technique is validated based on the study about Tohoku earthquake/tsunami event, and is implemented on the L-band ALOS/PALSAR data set. The analysis in this paper verifies that the proposed method can achieve accurate damage information.