Experimental results and analysis of sparse microwave imaging from spaceborne radar raw data

Abstract

Sparse microwave imaging is a novel radar framework aiming to bring revolutions to the microwave imaging according to the theory of sparse signal processing. As compressive sensing (CS) is introduced to synthetic aperture radar (SAR) imaging in recent years, the current SAR sparse imaging methods have shown their advantages over the traditional matched filtering methods. However, the requirement for these methods to process the compressed range data results in the increase of the hardware complexity. So the SAR sparse imaging method that directly uses the raw data is needed. This paper describes the method of SAR sparse imaging with raw data directly, presents the analysis of the signal-to-noise ratio (SNR) in the echo signal by combining the traditional radar equation with the compressive sensing theory, and provides the tests on 2-D simulated SAR data. The simulation results demonstrate the validity of the SNR analysis, and the good performance of the proposed method while a large percentage of the raw data is dropped. An experiment with RadarSat-1 raw data is also carried out to show the feasibility of processing the real SAR data via the method proposed in this paper. Our method is helpful for designing new SAR systems.