基于混合测量模式的压缩感知研究

Study on Compressed Sensing based on the Mixed Measuring Model

基于多层小波变换及压缩感知理论,提出了混合测量模式的压缩感知算法。该算法可以克服已有基于单层小波分解的压缩感知算法对于测量值点数的限制,使得算法更具有实用性。基本思想是对低频系数进行保留,仅对各层高频小波系数进行测量,即对小波系数采取混合测量模式。与经典压缩感知算法相比,在使用较少的测量值情况下,本文算法较好地提升了重构信号的质量。

Based on multilayer wavelet decomposition and compressed sensing theory,a compressed sensing algorithm with mixed measurement mode is proposed.This algorithm can overcome the limitation of existing compressed sensing algorithm based on single-layer wavelet decomposition on the number of measured points,and make the algorithm more practical.The basic idea is to retain the low-frequency coefficients and only measure the high-frequency wavelet coefficients of each layer,that is,the mixed measurement mode is adopted for the wavelet coefficients.Compared with the classical compressed sensing algorithm,in the case of using fewer measurements,the algorithm in this paper can better improve the quality of the reconstructed signal.