CN108304855A - More submarine characteristic signal blind source separation methods under a kind of marine environment - Google Patents

More submarine characteristic signal blind source separation methods under a kind of marine environment Download PDF

Info

Publication number
CN108304855A
CN108304855A CN201711267293.XA CN201711267293A CN108304855A CN 108304855 A CN108304855 A CN 108304855A CN 201711267293 A CN201711267293 A CN 201711267293A CN 108304855 A CN108304855 A CN 108304855A
Authority
CN
China
Prior art keywords
mask
cluster
clustering
signal
clusters
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201711267293.XA
Other languages
Chinese (zh)
Other versions
CN108304855B (en
Inventor
孙海信
王连生
成垦
沈艺珊
蔚然
王远旭
陈海兰
孙伟涛
苗永春
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xiamen University
Original Assignee
Xiamen University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xiamen University filed Critical Xiamen University
Priority to CN201711267293.XA priority Critical patent/CN108304855B/en
Publication of CN108304855A publication Critical patent/CN108304855A/en
Application granted granted Critical
Publication of CN108304855B publication Critical patent/CN108304855B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
  • Obtaining Desirable Characteristics In Audible-Bandwidth Transducers (AREA)

Abstract

The invention discloses more submarine characteristic signal blind source separation methods under a kind of marine environment, include the following steps：Acquire sample of signal；Sample vector based on the angles Hermitian in time-frequency domain and reference vector use clustering algorithm to be clustered to mixed signal sample to obtain multiple clustering clusters, using the membership function in clustering algorithm as mask；Cluster verification is carried out based on estimation number of clusters mesh；Mask is clustered into Q mask clustering cluster based on K means clustering algorithms；By mask estimation, Signal separator is carried out.The present invention can solve the problems, such as blind source separating in the case of owing fixed, and any mixed signal matrix or source position need not be estimated in mask estimation, the complexity of Signal separator is simplified, improves operation efficiency.

Description

More submarine characteristic signal blind source separation methods under a kind of marine environment
Technical field
The present invention relates to field of underwater acoustic signal processing, more particularly to more blind sources of submarine characteristic signal under a kind of marine environment Separation method and system.
Background technology
There are factors limitations, first blind separation algorithm to be used primarily in the transient state mixing of signal for current blind separation algorithm The blind separation algorithm of form, signal convolved mixtures form is relatively fewer, is broadly divided into following a few classes：1) it is based on eliminating signal point The Signal separator net of cross correlation, the network are solved the problems, such as using feedback circuit between amount.The algorithm have in practical applications compared with Big defect, in the case that scale between the signals differs greatly or hybrid matrix is morbid state, separating effect is poor or even nothing Method detaches signal.2) using nonlinear transfer function to output convert, such as based on InfoMax principle it is closely related most The method changed greatly.This algorithm the convergence speed is very slow, while separation matrix inverts and brings numerical value unstable.3) non-linear master point Amount analysis (PCA) algorithm, it is the popularization of linear principal component method.In addition, algorithm above is needing in mask estimation Estimate any mixed signal matrix or source position so that Signal separator process is relative complex, and operation efficiency is not high.
Invention content
The purpose of the present invention is to provide more submarine characteristic signal blind source separation methods under a kind of marine environment, can It solves the problems, such as blind source separating in the case of owing fixed, simplifies the complexity of Signal separator, improve operation efficiency.
To achieve the above object, the present invention uses following technical scheme：
More submarine characteristic signal blind source separation methods under a kind of marine environment, include the following steps：
S1, acquisition mixed signal sample；
S2, the sample vector based on the angles Hermitian in time-frequency domain and reference vector believe mixing using clustering algorithm Number sample is clustered to obtain multiple clustering clusters, using the membership function in clustering algorithm as mask；
S3, cluster verification is carried out based on estimation number of clusters mesh；
S4, mask is clustered by Q mask clustering cluster based on K- means clustering algorithms, is denoted as cq, q=1 ... Q, and it is full Sufficient Dq, the summation of q=1 ..., Q are minimum, DqIt is the total distance of mask and cluster barycenter in q-th of cluster, i.e.,：
Wherein,It is i-th of mask of kth frequency point, CqIt is q clusters cqBarycenter,It isWith cluster barycenter CqIt Between Pearson correlation coefficients, kstAnd kendThe beginning and end frequency point of the adjacent frequency group for cluster, i.e. the total number of frequency point Mesh is kend-kst+ 1,As distance metric, mask highly relevant in this way (smaller distance) will come from a cluster；
S5, pass through mask estimation, progress Signal separator.
Preferably, the clustering algorithm in step S2 uses K- mean clusters or Fuzzy C-Means Clustering Algorithm.
After adopting the above technical scheme, compared with the background technology, the present invention, having the following advantages that：
The present invention is based on the sample vector at the angles Hermitian and reference vectors, and mixed signal is had accumulated using clustering algorithm Sample detaches unknown signaling, can solve the problems, such as blind source separating in the case of owing fixed, need not estimate in mask estimation Any mixed signal matrix or source position, simplify the complexity of Signal separator, improve operation efficiency.
Description of the drawings
Fig. 1 is the flow diagram of the present invention；
Fig. 2 is the arrangement mode schematic diagram of source signal and receiver；
Fig. 3 is channel model schematic diagram；
Fig. 4 a give the original sound signals waveform of blue whale, and Fig. 4 b give the original sound signals waveform of humpback；
Fig. 5 a give the original sound spectrum figure of blue whale, and Fig. 5 b give the original sound spectrum figure of humpback；
Fig. 6 a give the mixed sound waveform of receiving terminal R1 captures, and Fig. 6 b give the mixed sound of receiving terminal R2 captures Waveform；
Fig. 7 a are the blue whale sound waveform isolated, and Fig. 7 b are the humpback sound waveform isolated；
Fig. 8 a are the blue whale sound audio spectrogram isolated, and Fig. 8 b are the humpback sound audio spectrogram isolated.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to the accompanying drawings and embodiments, right The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and It is not used in the restriction present invention.
Embodiment one
Referring to Fig. 1, the invention discloses more submarine characteristic signal blind source separation methods under a kind of marine environment, including Following steps：
S1, acquisition mixed signal sample.
S2, the sample vector and reference vector that Hermitian (hermitian matrix) angle is based in time-frequency domain, are calculated using cluster Method clusters mixed signal sample to obtain multiple clustering clusters, using the membership function in clustering algorithm as mask.Cluster Algorithm uses K- mean clusters or Fuzzy C-Means Clustering Algorithm.
S3, cluster verification is carried out based on estimation number of clusters mesh.
S4, mask is clustered by Q mask clustering cluster based on K- means clustering algorithms, is denoted as cq, q=1 ... Q, and it is full Sufficient Dq, the summation of q=1 ..., Q are minimum, DqIt is the total distance of mask and cluster barycenter in q-th of cluster, i.e.,：
Wherein,It is i-th of mask of kth frequency point, CqIt is q clusters cqBarycenter,It isWith cluster barycenter CqIt Between Pearson correlation coefficients, kstAnd kendThe beginning and end frequency point of the adjacent frequency group for cluster, i.e. the total number of frequency point Mesh is kend-kst+ 1,As distance metric, mask highly relevant in this way (smaller distance) will come from a cluster.
S5, pass through mask estimation, progress Signal separator.
Experimental evaluation
One, channel design
Two signal receiving terminals (R1 and R2), two acoustical signal transmitting terminals (S1 and S2) are simulated under water, and position is appointed Meaning.Wherein R1 and S1 at a distance of 3.5km, R2 and S2 at a distance of 4.0km, R1 and S2 at a distance of 4.5km, R2 and S1 at a distance of 4.7km, arrangement side Formula is as shown in Figure 2.
R1 and S1 forms channel h11, R1 and S2 and forms channel h12, R2 and S1 formation channel h21, R2 and S2 formation channel H22, four channel models are as shown in Figure 3.
Two, emulation experiment
The emulation experiment carries out on Matlab2014b platforms, and the processor of computer is AMD double-cores A6-4400M 2.7GHz inside saves as 4G.
The sound that emulation experiment uses be respectively blue whale, humpback cry, carried out variety classes whale sound respectively Separating experiment and both ends blue whale sound separating experiment.
Fig. 4 a give the original sound signals waveform of blue whale, and Fig. 4 b give the original sound signals waveform of humpback；Fig. 5 a The original sound spectrum figure of blue whale is given, Fig. 5 b give the original sound spectrum figure of humpback；Fig. 6 a give receiving terminal The mixed sound waveform of R1 captures, Fig. 6 b give the mixed sound waveform of receiving terminal R2 captures.
Three, separating resulting is analyzed
Mixed sound signal is detached by the blind source separation method of the present invention, the signal waveform isolated such as Fig. 7 institutes Show, wherein Fig. 7 a are the blue whale sound waveform isolated, and Fig. 7 b are the humpback sound waveform isolated.Respectively by Fig. 7 a and figure 4a, Fig. 7 b are compared with Fig. 4 b, it can be seen that blind source separation method of the invention can preferably be divided mixed signal From.
In order to further examine, Short Time Fourier Transform (STFT), obtained spectral image are done to the signal isolated As shown in figure 8, wherein Fig. 8 a are the blue whale sound audio spectrogram isolated, Fig. 8 b are the humpback sound audio spectrogram isolated.It will Fig. 8 a are compared with Fig. 5 a, and the characteristic energy of the two bottom is consistent；Fig. 8 b are compared with Fig. 5 b, the two bottom Characteristic energy is consistent.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto, Any one skilled in the art in the technical scope disclosed by the present invention, the change or replacement that can be readily occurred in, It should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with scope of the claims Subject to.

Claims (2)

1. more submarine characteristic signal blind source separation methods under a kind of marine environment, which is characterized in that include the following steps：
S1, acquisition mixed signal sample；
S2, the sample vector based on the angles Hermitian in time-frequency domain and reference vector, using clustering algorithm to mixed signal sample This is clustered to obtain multiple clustering clusters, using the membership function in clustering algorithm as mask；
S3, cluster verification is carried out based on estimation number of clusters mesh；
S4, mask is clustered by Q mask clustering cluster based on K- means clustering algorithms, is denoted as cq, q=1 ... Q, and meet Dq,q The summation of=1 ..., Q is minimum, DqIt is the total distance of mask and cluster barycenter in q-th of cluster, i.e.,：
Wherein,It is i-th of mask of kth frequency point, CqIt is q clusters cqBarycenter,It isWith cluster barycenter CqBetween Pearson correlation coefficients, kstAnd kendIt is the beginning and end frequency point of the adjacent frequency group for cluster, i.e. the total number of frequency point is kend-kst+ 1,As distance metric, mask highly relevant in this way (smaller distance) will come from a cluster；
S5, pass through mask estimation, progress Signal separator.
2. more submarine characteristic signal blind source separation methods under a kind of marine environment as described in claim 1, it is characterised in that： The clustering algorithm in step S2 uses K- mean clusters or Fuzzy C-Means Clustering Algorithm.
CN201711267293.XA 2017-12-05 2017-12-05 Multi-submarine characteristic signal blind source separation method in marine environment Active CN108304855B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711267293.XA CN108304855B (en) 2017-12-05 2017-12-05 Multi-submarine characteristic signal blind source separation method in marine environment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711267293.XA CN108304855B (en) 2017-12-05 2017-12-05 Multi-submarine characteristic signal blind source separation method in marine environment

Publications (2)

Publication Number Publication Date
CN108304855A true CN108304855A (en) 2018-07-20
CN108304855B CN108304855B (en) 2020-06-23

Family

ID=62870258

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711267293.XA Active CN108304855B (en) 2017-12-05 2017-12-05 Multi-submarine characteristic signal blind source separation method in marine environment

Country Status (1)

Country Link
CN (1) CN108304855B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112133321A (en) * 2020-09-23 2020-12-25 青岛科技大学 Underwater acoustic signal Gaussian/non-Gaussian noise suppression method based on blind source separation

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105282067A (en) * 2015-09-16 2016-01-27 长安大学 Complex field blind source separation method

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105282067A (en) * 2015-09-16 2016-01-27 长安大学 Complex field blind source separation method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
张良俊: "欠定盲源分离算法及其应用研究", 《中国优秀硕士学位论文全文数据库 (基础科学辑)》 *
郭鹿鸣: "基于盲源分离的心肺音信号分离方法研究与应用", 《中国优秀硕士学位论文全文数据库 (基础科学辑)》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112133321A (en) * 2020-09-23 2020-12-25 青岛科技大学 Underwater acoustic signal Gaussian/non-Gaussian noise suppression method based on blind source separation

Also Published As

Publication number Publication date
CN108304855B (en) 2020-06-23

Similar Documents

Publication Publication Date Title
Chen et al. Deep attractor network for single-microphone speaker separation
Huang et al. A blind channel identification-based two-stage approach to separation and dereverberation of speech signals in a reverberant environment
Kim et al. Blind source separation exploiting higher-order frequency dependencies
CN109830245A (en) A kind of more speaker's speech separating methods and system based on beam forming
WO2007100330A1 (en) Systems and methods for blind source signal separation
Shah et al. On the blind recovery of cardiac and respiratory sounds
Aichner et al. Time domain blind source separation of non-stationary convolved signals by utilizing geometric beamforming
CN108091345B (en) Double-ear voice separation method based on support vector machine
Kuo et al. Variational recurrent neural networks for speech separation
CN110544482B (en) Single-channel voice separation system
Mazur et al. An approach for solving the permutation problem of convolutive blind source separation based on statistical signal models
Han et al. DPCCN: Densely-connected pyramid complex convolutional network for robust speech separation and extraction
Jiang et al. An Improved Unsupervised Single‐Channel Speech Separation Algorithm for Processing Speech Sensor Signals
Dey et al. Single channel blind source separation based on variational mode decomposition and PCA
CN108304855A (en) More submarine characteristic signal blind source separation methods under a kind of marine environment
CN110265060B (en) Speaker number automatic detection method based on density clustering
Örnolfsson et al. Exploiting non-negative matrix factorization for binaural sound localization in the presence of directional interference
Jafari et al. Underdetermined blind source separation with fuzzy clustering for arbitrarily arranged sensors
CN114613384B (en) Deep learning-based multi-input voice signal beam forming information complementation method
Mitianoudis et al. Using beamforming in the audio source separation problem
Jafari et al. Sparse coding for convolutive blind audio source separation
CN103559886A (en) Speech signal enhancing method based on group sparse low-rank expression
Li et al. On loss functions for deep-learning based T60 estimation
Li et al. Expectation‐maximisation for speech source separation using convolutive transfer function
CN113241092A (en) Sound source separation method based on double-attention mechanism and multi-stage hybrid convolution network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant