SAR Data Compressing and Reconstructing Method Based on Compressed Sensing and LBG Algorithm

Aimed at the issue of how to compress huge SAR data apparently and obtain reconstructed results to complete SAR high resolution imaging for scene target, in this paper, a new approach combined with Compressed Sensing (CS) and LBG algorithm together is proposed. For SAR returned signal, firstly, according to CS theory, random Gauss noise matrix is designed as a measurement matrix to put forward data compressing. Secondly, Linde-Buzo-Gray (LBG) algorithm is employed to compress encode of every sample in order to complete diminishing data furthermore. What’s more, data reconstruction process still contains the two ordinal steps. One is decode process which is the inverse encode process of LBG algorithm. The other is original returned signal reconstruction by smooth L0 (SL0) algorithm according to CS theory. On the basis of that, the traditional Frequency Scaling (FS) algorithm is executed to achieve the final SAR image. The effectiveness of the proposed approach can be validated by simulation results.