Research on application of compressed sensing based on signal decomposition

Abstract

Based on compressed sensing (CS) theory and its classical OMP reconstruction algorithm, as to the problems that large storage of measurement matrix is needed when sampling and calculating the whole signal which has a large amount of data and that a huge amount of time is consumed when reconstructing it, we can decompose the signal into many subsignals through a certain way, and then measure and reconstruct the sub-signals. Finally, we can realize the processing of the whole segment signal. As above, the application of compressed sensing which based on signal decomposition can be called decomposed compressed sensing method in this paper. Currently, a commonly used decomposed compressed sensing method is segmented compressed sensing (SCS). For the problem that there are errors in processing certain types of sparse signals by SCS method, we propose a new decomposed compressed sensing in this paper. The simulation results show that compared with SCS, the new method can improve the reconstruction efficiency to a certain extent. At the same time, this method is applicable to any type of signal, which has wider application range.Index Terms-Compressed sensing (CS), decomposed CS, segmented CS, isometry sampling, orthogonal matching pursuit algorithm (OMP) ____________________________________ 978-1-4799-44-/14/$31.00 ©2014 IEEE