The Recovery Algorithm of Saturated Sar Raw Data Based on Compressed Sensing

Because of the unprediction of the scene scattering characteristic and the finite quantization bits, saturated data always exists. Saturation phenomenon leads to a non-linear distortion and interferes to the recognition of the target so that it affects the image quality. Especially when the scene scattering characteristic largely varies, it can generate false targets and degrade signal-to-noise ratio (SNR). Compressed sensing (CS), a non-linear reconstructed algorithm, is that samples in sub-Nyquist rate is used to recover the sparse signal with few non-zero elements. This paper proposes the recovery method based on the nonlinear characteristic of CS to recover the saturated part of the raw data to the unsaturation state and ensure the unsaturated parts maintain the original state. Simulation results validate the proposed method.