Introducing Partial Polarimetric Layers into a Curvelet-based Change Detection

Abstract

In this article a new change detection approach for polarimetric SAR data is presented. The further development of an already published change detection approach based on Curvelet transform for single polarized SAR images is combined with partial polarimetric layers of a reduced Huynen decomposition. The data used has been ac-quired by TerraSAR-X in the High Resolution Spotlight mode with two polarization channels: HH and VV (dual-pol). The motivation for using polarimetric information is that it helps to interpret changes indicated by the change detection algorithm, so that plant growth or crop on agricultural land as well as water level variations in rivers or even small scale changes in industrial plants like harbors get clearly visible. The Curvelet-based change detection turns again out to be a powerful tool for SAR image handling – image enhancement and change detec-tion – whereas the polarimetric information about the underlying scattering mechanisms delivers astonishing re-sults.