Skip to Main Content
Remote sensing is proving very useful for identifying damage and planning support activities after an earthquake has stricken. Radar sensors increasingly show their value as a tool for damage detection, due to their shape-sensitiveness, their extreme versatility and operability all weather conditions. The previous work of our research group, conducted on 1-m resolution Spotlight images produced by COSMO-SkyMed, has led to the discovery of a link between some selected texture measures, computed on radar maps over single blocks of an urban area, and the damage found in these neighborhoods. Texture-to-damage correlation was used to develop a SAR-based damage assessment method, but significant residual within-class variability makes estimations sometimes unreliable. Among the possible remedies the injection of physical vulnerability data into the model was suggested. This paper reports on a preliminary attempt in this direction.