Fast Encoding of Synthetic Aperture Radar Raw Data using Compressed Sensing

Authors:
Sujit Bhattacharya;Thomas Blumensath;Bernard Mulgrew;Mike Davies
Affiliations:
Institute of Digital Communication, Joint Research Institute for Signal and Image Processing, Edinburgh University, EH9 3JL. S.Bhattacharya@ed.ac.uk;Institute of Digital Communication, Joint Research Institute for Signal and Image Processing, Edinburgh University, EH9 3JL.;Institute of Digital Communication, Joint Research Institute for Signal and Image Processing, Edinburgh University, EH9 3JL.;Institute of Digital Communication, Joint Research Institute for Signal and Image Processing, Edinburgh University, EH9 3JL.
Venue:
SSP '07 Proceedings of the 2007 IEEE/SP 14th Workshop on Statistical Signal Processing
Year:
2007

Citing 0
Cited 4

On compressive sensing applied to radar

Signal Processing
Compressed sensing for synthetic aperture radar imaging

ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
SAR image reconstruction and autofocus by compressed sensing

Digital Signal Processing
SAR image Bayesian compressive sensing exploiting the interscale and intrascale dependencies in directional lifting wavelet transform domain

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Synthetic Aperture Radar (SAR) is active and coherent microwave high resolution imaging system, which has the capability to image in all weather and day-night conditions. SAR transmits chirp signals and the received echoes are sampled into In-phase (I) and Quadrature (Q) components, generally referred to as raw SAR data. The various modes of SAR coupled with the high resolution and wide swath requirements result in a huge amount of data, which will easily exceed the on-board storage and downlink bandwidth of a satellite. This paper addresses the compression of the raw SAR data by sampling the signal below Nyquist rate using ideas from Compressed Sensing (CS). Due to the low computational resources available onboard satellite, the idea is to use a simple encoder, with a 2D FFT and a random sampler. Decoding is then based on convex optimization or uses greedy algorithms such as Orthogonal Matching Pursuit (OMP).