The effects of input signal-to-noise ratio on compressive sensing SAR imaging

Abstract

Synthetic Aperture Radar (SAR) is an active and coherent microwave high resolution imaging system, which has the capability to image in all weather and day-or-night conditions. The high resolution required by various modes of SAR results in a huge amount of sampling data, which brings a demand for bigger storage. Recently, a novel concept based on Compressive Sensing (CS) theory asserts that an unknown sparse signal can be recovered exactly with an overwhelming probability even with highly sub-Nyquist-rate samples. In this paper, a new scheme for the test bed of CS based SAR imaging is proposed. Experimental results on some real raw SAR data reveal that there are some practical limitations on the use of CS based SAR imaging, especially for complex imaging scenes and systems with low Signal-to-Noise Ratio (SNR).