Multi-temporal Satellite Imagery for Flood Damage Assessment

Abstract

Assessment of fl ood prone region is carried out using multitemporal satellite imagery. The framework for precise assessment of damages caused by fl ood is presented using remote sensor data. We classify the framework into two categories: (1) Moderate satellite image processing using multi-temporal Moderate Resolution Imaging Spectroradiometer (MODIS) data, and (2) Optical and radar satellite image processing using Linear Imaging Self Scanning Sensor-III (LISS-III) and Synthetic Aperture Radar (SAR) data. In data set#1, spectral-spatial classifi cation is applied on before, during and after fl ood MODIS imagery. As MODIS is of moderate resolution hence two classes is considered namely water and nonwater region. In the case of data set#2 LISS-III (before fl ood) and SAR (during fl ood) data hierarchical classifi cation is applied. The SAR data is used to extract fl ooded and non-fl ooded regions where as LISS III data is used to classify various land cover regions. Further, the resultant images are overlaid to analyze the extent of the fl ood in individual land classes. We review the major development in processing these two fl ood assessment frameworks.