Damage assessment from VHR, spaceborne SAR imagery: including pre-event urban information

Inhalt:
In the context of risk assessment for natural disasters, radar sensors are more often used on the hazard side to monitor crustal displacement as an input to seismic models. However, it is also possible to take advantage of the shape-sensitiveness of radar data to map post-event damage to urban structures, a completely different type of information. Some methods have already been proposed, relying on amplitude patterns, on interferometric comparison between pre- and post-event data pairs, on statistical analysis of post-event amplitude data. It has already been shown how these latter can be improved slightly by including a priori information on the potentially damaged buildings such as a rough fragility estimation of buildings. In this paper, we introduce the idea of including a larger pool of pre-event information to specialize the statistical analysis carried out on the post-event dataset, in view of automated damage assessment. Specifically we consider including building density information as a potential additional input to the classifier, in accordance with the general consideration that such density impacts on both radar backscatter patterns and on combined fragility of buildings. The final goal of this research is to build an operational method capable to roughly assess the damage distribution and thus provide useful input to rescuers.