Fusion of gis and sar statistical features for earthquake damage mapping at the block scale

This work shows that multitemporal SAR data allows mapping earthquake damage in urban areas with an acceptable accuracy, once some ancillary information defining urban blocks is available. A statistical analysis of the parameters of the models representing backscatterer intensity or coherence values for each block may be used to discriminate between damaged and undamaged areas and, to some extent, to evaluate the damage. A comparison with a recently proposed supervised segmentation approach shows that the simpler, unsupervised methodology presented here can achieve comparable results