Abstract
This paper discusses a general methodology for post‐tsunami damage assessment and an automatic procedure able to distinguish between different kinds of damage on built‐up structures using very‐high‐resolution satellite data. The procedure for automatic detection of damaged built‐up structures was designed using a multi‐criteria approach fusing radiometric, textural, and morphological image features related to pre‐ and post‐disaster data detection. The proposed procedure shows good performance with an estimated overall accuracy equal to 93.97%. The best performances are estimated in the discrimination between non‐flooded and flooded built‐up structures and in the recognition of collapsed built‐up structures with debris in place. Problems of omission error were detected in the recognition of collapsed built‐up structures without debris in place as in the case of completed erased built‐up structures situated close to the shoreline.
Acknowledgements
The work presented here was in part funded by the Rapid Reaction Mechanism (RRM) budget line activated by European Commission DG RELEX after the Asian tsunami disaster. The work would also have been impossible without the collaboration of all the colleagues and friends belonging to the IPSC Support to External Security Unit, who greatly contributed in all phases from data collection to the analysis.