SNR analysis for SAR imaging from raw data via compressed sensing

Compressed sensing (CS) is a theory that guarantees the quality of reconstruction of a sparse signal from limited samples. The state-of-art CS radar imaging algorithms are all working on the compressed range data. They are feasible since matched filtering increases the signal to noise ratio (SNR). On the other hand, imaging from the raw data directly can reduce the hardware complexity of a radar system. However, according to the radar equation, the SNR in the radar echo signal is usually at a low level (e.g., −5dB). In this paper, we formulate a method of synthetic aperture radar (SAR) imaging from raw data via compressed sensing, and analyze SNR in the echo signal based on CS. Combining the traditional radar equation with the theory of compressed sensing, we provide an expression of SNR based on CS. The simulation results demonstrate the validity of our expression. An experiment of spaceborne stripmap SAR raw data is carried out, which shows the feasibility of SAR imaging from raw data via the method proposed in this paper.