Author(s)

Li Fen , Pan Xuefeng , Meng Gang

Corresponding Author

Li Fen

Abstract

The basic theories of signal sampling is the famous Nyquist sampling theorem, Nyquist sampling theorem pointed out that the bandwidth of the signal sampling rate reached more than two times, in order to accurately reconstruct the original signal by the sampling signal. However, with the increasing demand for information, the bandwidth of the signal carrying information more and more wide, so naturally raises a question: whether to ensure no loss of information, the sampling signal with a much lower rate than the Nyquist sampling theorem, but also can fully recover the signal. This paper describes the basic theory and core issues of compressed sensing, and introduces compression based reconstruction algorithms: base pursuit algorithm (BP algorithm) and orthogonal matching pursuit algorithm (OMP algorithm). Secondly, the basic principle of sparse representation is analyzed, the main methods of sparse representation are introduced, and the sparse representation method based on Gabor dictionary and its application in compressed sensing is emphasized.

Keywords

Compressed sensing, Signal reconstruction, Internet of things, Information sampling, Sparse matrix.