A Novel Random Sampling Method for Radar Image Compression

In this paper, we propose a process of the radar image data by Random sampling. The sampling rate is far lower than Nyquist-Shannon sampling theorem, which shows that image data can be reconstructed from an extremely small set of measurements than what is generally considered necessary. Compressive Sampling is considered for signals and images that are sparse in a wavelet basis. A low complexity compression method for high resolution image based on block partition in wavelet region is proposed. The image is partitioned into blocks in wavelet domain and then compressed separately with CS. The bit rates for each block is allocated according to the texture complexity of the block. The proposed method eliminates the "block effect" caused by traditional block partition in pixel domain, and solves the problem caused by traditional block partition that the areas with simple texture are good in reconstruction quality while those with complicated texture are too poor to be used because of the uneven distribution of texture complexity. The experimental results show that the compression performance of the proposed method is quite similar to the results obtained by global compression. This compression method is particularly suitable for the high resolution remote sensing image which sparse in a wavelet domain.