A New Method for Selecting Coherent Scatterers Based on Their Polarimetric Signature

Persistent Scatterers (PS) technique is a useful modification of InSAR that limits interferometric processing only to those pixels that behave consistently over a long period of time. Usually these pixels are selected by setting a threshold on amplitude dispersion or mean coherence. This approach requires a large number of images in order to select PS with high confidence which can be a significant limitation particularly at early stages of data acquisition. We propose another approach for selecting coherent scatterers based on their polarimetric properties. This approach requires only a few quad-pole SAR images. By performing polarimetric analysis we are able to select only those pixels that behave consistently over a studied period and apply standard InSAR processing if network of selected pixels is dense or point processing similar to PS if network of selected scatterers is sparse. As an example we present results of this technique utilizing quad-pole RADARSAT-2 data over San Francisco. Introduction Differential Synthetic Aperture Radar Interferometry (DInSAR) is a technique for measuring ground deformation with a high spatial resolution and accuracy over a large area [Massonnet and Feigl, 1998; Rosen et al., 2000]. The differential interferogram is calculated from two Synthetic Aperture Radar (SAR) images acquired by the same or a similar satellite at two different times and displays the line-of-sight (LOS) deformation that occurred between the two acquisitions. The quality of an interferogram is described by its coherence, which is the magnitude of a cross-correlation coefficient calculated for a master-slave pair of images. Coherence depends on a variety of parameters, including SAR wavelength, incidence angle, perpendicular baseline and land cover. In a densely vegetated environment coherence generally is low, which prevents the formation of a high-quality interferogram and phase unwrapping. For these land cover conditions, Persistent Scatterers (PS) interferometry technique is used instead [Ferretti et al., 2000; 2001]. The PS approach is based on a selection of pixels that are dominated by a single strong scatterer within the pixel and which are consistent over a long period of time. These pixels are usually selected based on their amplitude or amplitude dispersion. The standard PS approach has also a few disadvantages. In order to confidently select PS pixels it is necessary to have a large number of images. In general, selection of PS pixels with a high degree of accuracy is possible only when at least thirty SAR images are available [Ferretti et al., 2001]. This is not always possible or desirable. Not only is it often difficult or expensive to acquire large numbers of images, but that many images inevitably span large time periods, on the order of years to decades. Hazard estimates for volcanic or seismic monitoring, for example, often requires the production of time series and deformation rates on short time frames, ie. after the acquisition of only a few images. In addition, the selection of PS pixels is limited to those pixels that have very high amplitude, low standard deviation or satisfy both conditions. In some regions the network of such PS pixels is very sparse, which limits the processing of such interferograms, particularly phase unwrapping. In this paper we demonstrate a technique that is able to select a larger number of Persistent Scatterers by analyzing their polarization phase difference [Samsonov and Tiampo, submitted to RSE; Evans et al.,