SAR image Bayesian compressive sensing exploiting the interscale and intrascale dependencies in directional lifting wavelet transform domain

Compressive Sensing (CS) provides a new solution to reduce the huge amount of data for the transmission and storage of high resolution synthetic aperture radar (SAR) images. To improve the CS performance, in this work we propose directional lifting wavelet transform (DLWT) as a sparse representation for SAR image CS. Then a Bayesian-based SAR image CS reconstruction algorithm in DLWT domain is proposed. To further improve the reconstruction performance, an accurate prior probability model is proposed which fully exploits interscale attenuation and intrascale directional clustering properties of the DLWT coefficients; and a Bayesian inference via Markov Chain Monte Carlo (MCMC) sampling is employed to recover the image[U+05F3]s wavelet coefficients and SAR image. Experimental results show that the proposed DLWT Tree-Direction-Clustering Compressive Sensing (DLWT-TDC-CS) can achieve the best reconstruction performance at sampling rates from 0.5 to 0.9 compared with various state-of-the-art CS reconstruction algorithms. DLWT-based CS reconstruction outperforms DWT-based CS reconstruction due to the improved sparse representation.