Assessment of the Change Detection Procedure Dedicated to Flood Monitoring Using Envisat Wide Swath Mode Data

The aim of this paper is to assess the change detector r specifically elaborated to analyse ENVISAT ASAR Wide Swath Mode (WSM) data pairs for plain flood changes detection. The algorithm relies on an enhanced ratio for strong changes over large homogeneous and flat areas combined with a ratio calculated from the two raw images. The change detection assessment was conducted on 14 image pairs exploiting the ENVISAT WSM database set up over the Poyang Lake within the framework of the Flood DRAGON Project. The best results were obtained by performing a supervised classification on a colour composite combining the change detector r, the before and after images. Change extraction accuracy reaches 86.3% and is very robust despite the differences in data polarization combinations. Mainly focused on plain flood changes, the proposed change detection procedure is also sensitive to changes relative to intensive shipping, rice cycles and wetland vegetation growth.