Short-range ground-based synthetic aperture radar imaging: performance comparison between frequency-wavenumber migration and back-projection algorithms

Abstract

Two popular synthetic aperture radar (SAR) reconstruction algorithms, namely the back-projection (BP) and the frequency wavenumber (ω−k ) algorithms, were tested and compared against each other, especially for their use in ground-based (GB) SAR applications directed to foreign object debris removal. For this purpose, an experimental setup in a semi-anechoic chamber room was accomplished to obtain near-field SAR images of objects on the ground. Then, the 90 to 95 GHz scattering data were acquired by using a stepped frequency continuous-wave radar operation. The performances of the setup and the imaging algorithms were then assessed by exploiting various metrics including point spread function, signal-to-clutter ratio, integrated side-lobe ratio, and computational complexity. Results demonstrate that although both algorithms produce almost accurate images of targets, the BP algorithm is shown to be superior to the ω−k algorithm due to its some inherent advantages specifically suited for short-range GB-SAR applications.