Processing Polarimetric SAR Time Series over urban Areas with binary partition Trees

Abstract

During the last decade, Synthetic Aperture Radar (SAR) sensors have demonstrated its usefullness for monitoring the surface of the Earth. They are able to extract information related to the scene reflectivity at microwave frequencies [1]. Additionally, since they are active sensors, they present some important advantages with respect to other sensors, as they are almost independent of the weather conditions and the day-night cycles. Furthermore, in Polarimetric SAR (PolSAR) different polarizations are employed for transmitted and received radar echoes, and some additional features may be extracted from the target and its geometry [2]. It is for these reasons that SAR systems mounted on a space-borne system can provide very valuable time series datasets. These datasets are very useful to monitor the scene and, specially, its temporal evolution. Over urban areas, PolSAR time series may be effective for monitoring the different urban activities and changes. Polarimetry is particularly convenient for urban areas, since most human made structures present a strong radar reflectivity and, additionally, its sensitivity to the target structure is very useful for monitoring the urban evolution.