Babak Mansouri - Assistant Professor, International Institute of Earthquake Engineering and Seismology, Tehran
Masanobu Shinozuka - Distinguished Professor and Chairman, University of California Irvine, Dept. of Civil & Environmental Eng., California

This research overviews the systematic progress of change and damage detection capabilities in urban areas using satellite SAR data. Marmara earthquake occurred on August 17, 1999 with a magnitude of 7.4 and caused extensive urban damage in the northwestern region of Turkey. Town of Seymen is being focused for this earthquake. The city extensively suffered building damage with dominant mode of pancaking failure. The approach is based on rapid and postprocessing of SAR data using image processing and change detection algorithms, especially correlation and coherence map generation applied to before and after SAR (ERS1&2 satellites) complex images. Implemented algorithms show valuable means for change detection and can indicate the location of building damages in regional sense. For the purpose of change detection in Bam after the December 26, 2003 earthquake, improved algorithms were applied to the more sophisticated Envisat data. In order to improve the sensitivity for change detection and to eliminate false alarm in change detection, Radar Cross Section (RCS) calibration curve is used in the algorithm and applied on the SAR cross-power difference map. The high resolution Quickbird image of Bam is used for creating urban zones based on building orientations. RCS curve use SAR-Target orientation information and calibration mask is generated and applied to the preliminary change map. Finally, this improved map is compared with the results provided by the field.

SAR, Damage Detection, Urban Damage, Correlation, Coherence, Cross-power, RCS