Implementation of Backprojection Algorithm for Synthetic Aperture Radar Image Processing on Low-Cost Hardware Platform

Abstract

This work presents the implementation of back-projection algorithm for Synthetic-Aperture Radar (SAR) signals on a low-cost, small, lightweight, and low-power consumption platform: Raspberry Pi. The algorithm is implemented with GNU Octave open-source software and the performance was tested on Raspberry Pi 3B and 4 hardware. For performance comparison, a single-threaded baseline implementation of back-projection is created and then modified to run on several threads on an available multicore processor. Executing a single-threaded code Raspberry PI is too slow for real-time imaging. However, the parallelized version shows computation improvement over the baseline version. We include a discussion of parallel implementation on a single Pi using Octave’s parallel package. This study contributes to the understanding of implementing SAR image processing on affordable single-board platforms with constrained computing resources.