Font Size: a A A

Research On Denoising Method Of Low Frequency Underwater Acoustic Signal Based On Sparse Decomposition

Posted on:2022-03-21Degree:MasterType:Thesis
Country:ChinaCandidate:Y W WuFull Text:PDF
GTID:2518306527950499Subject:Software engineering
Abstract/Summary:PDF Full Text Request
The 21 st century is the century of the ocean.Many scholars have launched various marine scientific researches around the world.Among them,the actual application of underwater target detection,navigation,communication,etc.All these researches require accurate underwater acoustic signals,which makes the denoising of underwater acoustic signals a top priority in marine scientific research.Aiming at this problem,this paper conducts in-depth research on traditional sparse decomposition algorithms and dictionary learning algorithms,and uses them as the research basis.Combined with the difficult problem of long pulse signals,we improve the traditional sparse decomposition algorithms,and solve the limitations of traditional sparse decomposition algorithms.In this essay,the denoising processing of noisy underwater acoustic signals is carried out in the shallow sea environment.The main research contents include:(1)Choosing an appropriate signal and noise model.In this paper,we select the normal mode model to calculate the sound field according to the characteristics of the underwater acoustic signal,then we obtain the sound pressure value of the underwater acoustic signal.We choose a non-Gaussian noise model to simulate the ocean noise environment more realistically.(2)Under the above conditions,we study the traditional sparse decomposition algorithm.We improve the matching pursuit(MP)algorithm and the orthogonal matching pursuit(OMP)algorithm for the problem of the excessive number of atoms in the over-complete dictionary.And we use one of the algorithms respectively to perform denoising processing on the received signals under different signal-to-noise ratios(SNRs).(3)We study the method of optimal directions(MOD)and the K-Singular Value Decomposition(K-SVD).Based on the OMP algorithm,two dictionary learning algorithms are added respectively to carry out the denoising process under different sea conditions.(4)Aiming at the difficult problem of the measured long pulse signal,we add pulse compression technology to compress it,solving the limitation of the traditional sparse decomposition algorithm on the long signal,and reduce the calculation amount of the algorithm while improving the SNR of the reconstructed signal.In different noise environments and in the processing of different forms of signals,the improved methods have all achieved good denoising effects.The results of simulation and actual sea trial data show that the improved methods are efficient in removing additive noise from the received signals and it breaks the limitation of traditional sparse decomposition algorithm for processing long pulse signals.
Keywords/Search Tags:underwater acoustic signals, sparse decomposition, dictionary learning, pulse compression
PDF Full Text Request
Related items