Abstract
This paper discusses using filterbank approaches for complex spectral estimation and Synthetic Aperture Radar (SAR) imaging. Specifically, we present a method, referred to as APES (Amplitude and Phase EStimation), for complex spectral estimation and SAR imaging. Two different derivations of the APES method are given, of which the first mimics the Maximum Likelihood (ML) methodology, and the second gives rise to a general class of MAtched-FIlterbank (MAFI) approaches to spectral estimation. We show that the widely used Capon estimator, originally derived in a different framework, also belongs to the general class of the MAFI approach. The MAFI interpretation of Capon and APES, along with some additional analyses, provides interesting insights into the properties of and the relation between them. Numerical studies that include general spectral estimation exercises and SAR imaging using experimental data are presented to demonstrate the performances of these methods.
