Reconstruction method of sparse signal

Abstract

The invention discloses a reconstruction method of sparse signal, mainly solving the problem of low rate of raw sparse signal reconstructed from observation vectors. Utilizing the decomposability of aconstrained object function, the method decomposes the problem of optimizing the constrained object function into a series of small constrained object functions for optimization to improve the reconstruction rate, comprising a raw data acquisition part, a raw sparse signal reconstruction part and a reconstruction accuracy evaluation part, wherein the raw data acquisition part comprises raw sparsesignal generation, matrix observation and vector observation; the raw sparse signal reconstruction part mainly comprises setting unconstrained object function, deducing constrained object function, decomposing the constrained object function by a sequential minimal optimization method, calculating reconstruction signal and debiasing reconstruction signal; the reconstruction accuracy evaluation isto compare the magnitude of the error of mean square. The reconstruction method can improve the reconstruction rate on the premise of ensuring reconstruction accuracy rate and be used for solving theproblem of sparse signal reconstruction in the fields of compressed sensing and the like.

Description

technical field [0001] The invention belongs to the technical field of signal processing, in particular relates to sparse signal reconstruction, which can be used in compressed sensing or the opposite problem. Background technique [0002] With the development of the digital information age, the analogization of information sources and the digitization of information processing tools are becoming more and more dominant, and signal sampling is the bridge between the two. The Fourier transform and the Nyquist sampling theorem show that to recover a signal from a discrete signal without distortion, the sampling rate must be twice the signal bandwidth. However, with the increase of people's demand for information, the bandwidth of the signal carrying information is getting wider and wider. According to this theory-based signal processing framework, the required acquisition rate and processing speed are getting higher and higher. It is also getting bigger and bigger, which bring...

AI Technical Summary

Problems solved by technology

l1_ls is a method based on the conjugate gradient descent algorithm. Every cycle needs to find the direction of descent, which wastes time.

Although the OMP method saves a lot of time, its robustness to noise is too low for inconsistencies in compressed sensing

GPSR and state-of-the-art waste time when there are many loops

Images