Revisiting the South Indian floods of 2015 with Sentinel-1 data

Earth monitoring from space has been made possible with the increase in number of Synthetic Aperture Radar (SAR) sensors and the availability of data of the same scene with frequent revisiting time. The flood mapping using SAR sensors provide immediate response map. In this paper, the flood mapping with Sentinel-1 data using Mahalanobis classifier on preprocessed SAR data and texture features exploited from co-occurrence pixels to distinguish flooded and non-flooded area is introduced. The accuracy of the introduced technique is estimated using the available ground truth. The 2015 South India floods that devastated the metropolitan city of Chennai is revisited using the SAR acquisitions and the corresponding flood map using the introduced technique is derived and the results are analyzed.