Improved Disaster Management through Post-earthquake Building Damage Assessment Using Multitemporal Satellite Imagery

Remote sensing technology is playing an increasingly important role in post-disaster decision support. Organizations including the Earthquake Engineering Research Institute (EERI) and Multidisciplinary Center for Earthquake Engineering Research (MCEER) recognize that following major earthquakes, satellite imagery can bring significant benefits to response and recovery efforts, through urban damage assessment. Following an introductory review of prior research undertaken in the field, this paper presents methodological techniques for determining the location and severity of post-earthquake building damage. The damage detection algorithms are based on the comparative analysis of a multitemporal sequence of optical or SAR images, acquired before and after the event. Following initial pre-processing to geo-reference and co-register the imagery, damage is detected in terms of spectral changes between the scenes. Depending on its spatial resolution, image processing techniques including edge detection and texture analysis may then be used to isolate features of interest within the coverage. Change is subsequently quantified with a range of arithmetic operators. Damage location is visualized using damage maps, and severity expressed graphically as damage profiles. The efficacy of these analytical procedures is demonstrated with respect to the Marmara (Turkey) earthquake of August 17th 1999, for which 10m panchromatic SPOT 4 and 20m SAR ERS coverages were available. Preliminary results are also presented for the May 21st, 2003 Boumerdes earthquake in northern Algeria, where 60cm Quickbird imagery was acquired. In this latter case, the change detection algorithms offer a quick-look region-wide damage assessment, providing the focus for more detailed visual inspection of building damage on a per-structure basis. The development of damage profiles for Boumerdes is an ongoing research topic. In future earthquakes, these techniques could guide the work of field reconnaissance teams; support the prioritization of relief efforts; direct search and rescue teams to victims; facilitate loss estimation; and help determine whether the situation warrants international aid. * Corresponding author.