Parcel-based Damage Detection using VHR Optical Data

Optical remote sensing has gained major technology breakthroughs and outstanding merits in monitoring the urban areas in disaster conditions. However, any change detection procedure must be checked for its level of sensitivities to different aspects of the phenomenon, materials and aspects of changes. So far, pixel-based and object-based image processing algorithms have been well received and utilized for major disasters around the world. These results are more useful in rapid loss estimation schemes right after a major disastrous event but baring in mind that these schemes are relatively imprecise. Considering the fact that using urban parcel information improves the quality of the damage detection results, this research focuses on incorporating a city parcel database with a novel optical change detection algorithm. The parcel information is developed from the city CAD files (for BAM). The parcels are extracted from stereo aerial photos and modified with a manual process using very high resolution optical data in GIS. Our optical change detection algorithm takes into account different spectral and spatial features of the panchromatic band of Quickbird satellite data. A fuzzy logic methodology is then applied at the end of the process to identify the change s. The final results are then compared with direct visual damage observation of the building using VHR data.