Super-Resolution Processing for Polarimetric Synthetic Aperture Radar Tomography

In polarimetric SAR interferometry, fully-polarimetric measurements are obtained along multiple interferometric tracks and used to generate three-dimensional tomographic images of a scene of interest. The objective is to utilize the polarimetric diversity of the targets to separate multiple targets within a single pixel. In this study, a new technique known as Vector Linear Prediction (VLP) is developed and used to perform polarimetric aperture extrapolation with measurements from a single baseline SAR interferometer. The VLP method extends scalar, linear prediction-based extrapolation to coherent, multi-channel data such as polarimetric SAR interferometric measurements. The interferometric phase estimates, which reflect the elevation of the scattering centers, are obtained for both the ground and vegetation components. The phase difference between the two components is shown to be consistent with the vegetation cover in the region under study. In the future, this technique will be applied to multi-baseline SAR measurements to produce tomographic images of the scenes being imaged.