Mapping Damaged Buildings through Simulation and Change Detection of Shadows using LiDAR and Multispectral Data

摘要

A practical processing framework for EO-based detection of building damage in dense urban areas is proposed basedon pre- and post-event shadow differencing. The basic data set used for the detection of damaged buildings includesLiDAR and multispectral images with high spatial resolution. The typical building damage types after a majorearthquake, such as height-reduced, overturn collapse and inclination, have been considered in this study. Through ascenario case study based on simulations of both building damage and shadow, understandings of the relationshipbetween shadow and building damage are improved for real-time response practices.