Turbo-like Iterative Thresholding for SAR image recovery from compressed measurements

Compressive sensing (CS) has attracted many researchers since it offers a novel paradigm that one can acquire sparse signals at a sub-Nyquist rate without information losses. In J. Ma, April 2009, S. Bhattacharya et al, Aug 2007, and G. Rilling et al, 2009, the authors have presented some schemes for CS application on remote sensing imaging, some of which are related to SAR. CS remote sensing imaging includes two steps: on-board encoding imaging and off-line decoding recovery. Based on the on-board encoding imaging scheme proposed in J. Ma, April 2009, this paper focuses on the off-line decoding recovery algorithm. We proposed a turbo-like iterative residual thresholding algorithm (RTIT) to decode the compressed SAR data with approximately sparse property. The experimental results show that it outperforms the state-of-the-art iterative thresholding algorithm (IT).