Efficient Wavenumber Domain Focusing for Ground-Based SAR

Abstract

We present an efficient focusing algorithm for synthetic aperture radar (SAR) data acquired in short bursts by a geometry with a large range spread, i.e., the case of ground-based (GB) SAR. The usual approach for focusing GB SAR data is the polar formatting algorithm, whose computational complexity, however, is quite relevant due to the nonseparability of the kernel. In this letter, we introduce a different format for the focused data, namely, the range-angular domain. Such format keeps the benefit of the polar format that samples data close to the resolution, but allows for the design of a separable kernel in the wavenumber domain. The proposed kernel exploits a modified Stolt interpolation and an efficient space-varying resampling. Results of processing on both simulated and real data are presented.