The Optimization Design of Measurement Matrix Based on KSVD-ETF

Compressive sensing, a novel signal acquisition method, is a joint sensing-compression process which re- quires a small number of measurements to reconstruct signal. Measurement matrix, a very important part in compres- sive sensing, directly affects the adaptive sparsity, the required number of measurements and the reconstruct perform- ance of the signal. In order to decrease the mutual coherence between the measurement matrix and sparse transformed matrix and improve the quality of reconstruction, this paper addresses the joint optimization between measurement ma- trix and sparse dictionary based on the KSVD-ETF. While optimizing the measurement matrix by ETF, we use the KSVD method to update the dictionary. The PSNR of the reconstructed signal is improved with the optimized meas- urement matrix from the experimental results, indicating that this method of optimizing the measurement matrix has certain advantages in the effect of reconstruction.