The Parallelization of Synthetic Aperture Radar (SAR) Image Focusing Algorithms on GPU

Abstract

The increased demand for higher resolution and detailed SAR imaging builds up a pressure on the processing power of the existing systems for real time or near real time processing. Exploitation of GPU processing power could suffice the increasing demands in processing. The processing of initial SAR systems was based on the principles of Fourier Optics. Lenses provided a real time two-dimensional Fourier transform of the data This document comprises results and analysis of parallelizing Range Doppler and Chirp scaling algorithms for SAR imaging and comparison of computational time over traditional CPU and GPU platform. The results shows that RDA in its essence gives better speed-up than CSA basically due to its less complex manipulations.