Earthquake Damage Detection in Built Environment: An Object-Oriented Approach Using Radar Imagery

Abstract

The feasibility to develop a rapid urban seismic damage detection procedure utilizing satellite radar imagery is investigated. In direct observation of city-wide building loss, remote sensing damage detection techniques have shown merits in rapid damage detection in urban areas. Remote sensing technology has the capability of extracting buildings in urban scenes. Comparing before and after images and benefiting from image processing techniques, it is possible to detect the extent of high hit zones. In detecting post-earthquake damages to buildings and in order to reduce detection errors and for minimizing the false alarms, it seems positive logic to apply the change detection algorithms only to the patches that correspond exactly to building footprints. For this purpose, urban database revealing the 3D reconstruction of the city is developed using parcels records. The parcels are extracted from aerial photos (stereography processing) then complemented and updated using VHR (very high resolution) optical satellite image (i.e. Quickbird imagery). The change detection algorithm and the calibration modeling are applied to before and after EnviSat SAR images considering only the building layer. The methodology is applied to the city of Bam and the associated building damages of its 2003 earthquake were emphasized. Results were compared with an independent direct visual damage interpretation using a VHR optical image.