Large scale change detection techniques dedicated to flood monitoring using ENVISAT wide swath mode data

Semi-automatic flood extraction procedures based on change detection techniques appear particularly adapted to plain flood monitoring and mapping. The change detector was specifically elaborated to analyse ENVISAT ASAR wide swath mode data pairs, which appear very well adapted to flood monitoring over wide areas. The algorithm is based on two analysis levels: an enhanced ratio for strong changes over large homogeneous and flat areas combined with a ratio calculated from the two raw images which aims to keep the raw data's thematic precision. Slope and aspect effects are also eliminated by the use of a digital elevation model during the processing. This change detection analysis was performed on 35 data pairs acquired within the framework of the flood DRAGON project. The first set of results is very promising and robust using HH polarization. Change detection between data with W polarization or with different polarization (HH versus W) has to be fully validated but results are promising.