Using omega‐ K algorithm for bistatic synthetic aperture radar image formation based on modified extended Loffeld's bistatic formula

Abstract

Bistatic synthetic aperture radar (BSAR) as a way of Earth remote sensing has been developed considerably in recent years, both theoretically and practically due to its unsubstitutable services. BSAR frequency-domain processing algorithms are efficient ways of image formation in comparison to ideal two-dimensional matched filtering and to relatively accurate and time-consuming time-domain algorithms. Among these frequency-domain algorithms, omega-K is the most precise. The starting and key step of frequency-domain algorithms is the derivation of bistatic spectrum. Recently, a new bistatic spectrum is reported, which is probably the latest and the most modified version of Loffeld's bistaic formula (LBF), maintaining its accuracy even in azimuth-variant configurations with high squint angles. So far, this spectrum has only been used within range-Doppler algorithm to process BSAR data. The authors investigate the possibility and results of applying this modified version of LBF as a basis for omega-K algorithm. Two approaches, based on Stolt interpolation and inverse scaled Fourier transform, are examined and their effectiveness in general azimuth-variant geometry is validated through several simulations. The proposed implementations show higher performance in terms of image quality measurements as compared to extended LBF-based implementations.