A Fully-Polarized Unitary MUSIC for Polarimetric SAR Tomography

Abstract

In this paper, a fully-polarized unitary multiple signal classification (MUSIC) for polarimetric SAR tomography is proposed. Compared with the classical MUSIC, unitary MUSIC utilizes the conjugate of complex sampling data to achieve higher resolution when the polarimetric SAR tomography has a small number of baselines. The combination of unitary MUSIC and full polarization permits to further reduce the noise of sample covariance matrix on the basis of multilooking consisting of averaging pixels. By making full use of the target s fully-polarized information, this algorithm not only improves the resolution in the elevation dimension but also obtains the accurate polarimetric scattering matrix (PSM) of a target. Simulation results show that this algorithm has better performance than the popular distributed compressed sensing (DCS).