POST-DISASTER DAMAGE ASSESSMENT THROUGH COHERENT CHANGE DETECTION ON SAR IMAGERY
- 1Department of Electronics, Information and Bioengineering, Politecnico di Milano, Milan, Italy
- 2Interuniversity Department of Regional and Urban Studies and Planning, Politecnico di Torino, Turin, Italy
- 3Department of Electronics and Telecommunications, Politecnico di Torino, Turin, Italy
- 4Department of Civil And Environmental Engineering, Politecnico di Milano, Milan, Italy
Keywords: Coherent Change Detection, Synthetic Aperture Radar, Automatic post-disaster damage assessment, Earthquake
Abstract. Damage assessment is a fundamental step to support emergency response and recovery activities in a post-earthquake scenario. In recent years, UAVs and satellite optical imagery was applied to assess major structural damages before technicians could reach the areas affected by the earthquake. However, bad weather conditions may harm the quality of these optical assessments, thus limiting the practical applicability of these techniques. In this paper, the application of Synthetic Aperture Radar (SAR) imagery is investigated and a novel approach to SAR-based damage assessment is presented. Coherent Change Detection (CCD) algorithms on multiple interferometrically pre-processed SAR images of the area affected by the seismic event are exploited to automatically detect potential damages to buildings and other physical structures. As a case study, the 2016 Central Italy earthquake involving the cities of Amatrice and Accumoli was selected. The main contribution of the research outlined above is the integration of a complex process, requiring the coordination of a variety of methods and tools, into a unitary framework, which allows end-to-end application of the approach from SAR data pre-processing to result visualization in a Geographic Information System (GIS). A prototype of this pipeline was implemented, and the outcomes of this methodology were validated through an extended comparison with traditional damage assessment maps, created through photo-interpretation of high resolution aerial imagery. The results indicate that the proposed methodology is able to perform damage detection with a good level of accuracy, as most of the detected points of change are concentrated around highly damaged buildings.