Accelaerated SAR image generation on GPGPU platform

Image generation from satellite-borne synthetic aperture radar (SAR) video data requires numerical computations of the order of gigaflops (GFLOPs). Current parallel processing schemes distribute computation load among a number of parallel processing elements (PEs) for improving efficiency in computation. In order to handle the huge computational task for real-time image generation, we report here the use of general purpose graphics processing units (GPGPUs) that have shown excellent results for data-parallel kinds of operations. In this paper, we describe the architecture, and highlight the steps needed to extract parallelism in implementation of the algorithm for accelerated SAR image generation.