CN109413543A - A kind of source extraction method, system and storage medium - Google Patents
A kind of source extraction method, system and storage medium Download PDFInfo
- Publication number
- CN109413543A CN109413543A CN201710698651.6A CN201710698651A CN109413543A CN 109413543 A CN109413543 A CN 109413543A CN 201710698651 A CN201710698651 A CN 201710698651A CN 109413543 A CN109413543 A CN 109413543A
- Authority
- CN
- China
- Prior art keywords
- signal
- input signal
- input
- interference
- target jamming
- 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
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R3/00—Circuits for transducers, loudspeakers or microphones
- H04R3/005—Circuits for transducers, loudspeakers or microphones for combining the signals of two or more microphones
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R3/00—Circuits for transducers, loudspeakers or microphones
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0272—Voice signal separating
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R25/00—Deaf-aid sets, i.e. electro-acoustic or electro-mechanical hearing aids; Electric tinnitus maskers providing an auditory perception
- H04R25/40—Arrangements for obtaining a desired directivity characteristic
- H04R25/407—Circuits for combining signals of a plurality of transducers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R1/00—Details of transducers, loudspeakers or microphones
- H04R1/20—Arrangements for obtaining desired frequency or directional characteristics
- H04R1/32—Arrangements for obtaining desired frequency or directional characteristics for obtaining desired directional characteristic only
- H04R1/40—Arrangements for obtaining desired frequency or directional characteristics for obtaining desired directional characteristic only by combining a number of identical transducers
- H04R1/406—Arrangements for obtaining desired frequency or directional characteristics for obtaining desired directional characteristic only by combining a number of identical transducers microphones
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R2225/00—Details of deaf aids covered by H04R25/00, not provided for in any of its subgroups
- H04R2225/43—Signal processing in hearing aids to enhance the speech intelligibility
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R2430/00—Signal processing covered by H04R, not provided for in its groups
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R2430/00—Signal processing covered by H04R, not provided for in its groups
- H04R2430/20—Processing of the output signals of the acoustic transducers of an array for obtaining a desired directivity characteristic
- H04R2430/23—Direction finding using a sum-delay beam-former
Abstract
The disclosure provides a kind of method, system and storage medium that target jamming signal is persistently extracted from selected signal.The described method includes: acquisition two-way or multichannel input signal, every road input signal contain target jamming signal；Improve the independence of the input signal；Calculate the resulting coefficient matrix of independence for promoting the input signal；Synchronize each pair of or every group of input signal；Input signal after synchronizing is separated into target jamming signal and useful signal；Intelligent selection output signal.
Description
Technical field
This disclosure relates to signal processing technology field more particularly to a kind of for extracting interference signal from mixed signal
Signal processing technology.
Background technique
In current demand signal processing and big data field, the measurement of observation signal often by the interference of garbage signal, because
This, how to improve the signal-to-noise ratio of surveyed observation signal is a huge challenge.Same problem also appears in recording (such as film studio
Recording, hearing aid, 360 audio devices), biomedical applications (such as brain wave record, Brian Imaging) and remote sensing (such as radar signal,
Echolocation) etc. fields.Eliminating the most common method of this kind of interference signal is the filter using analog or digital form.But
It is that useful signal and interference signal often share a frequency range, and filter is difficult to separate them.
Current isolation technics operates hearing devices mainly by way of selective control signal specific gravity, and emphasis focuses
The design factor matrix how more effectively, or using a directional microphone and an omnidirectional microphones combination come
Enhance the clarity of voice, but traditional Independent Component Analysis (ICA) is unable to reach ideal effect, to interference signal
Removal effect is undesirable, and destroys the accuracy of ICA.
Therefore, currently it is badly in need of a kind of technology that can effectively separate useful signal and interference signal.
Disclosure
The problem of existing in view of the above technology, the disclosure is creatively by doing the sides such as Domain Synchronousization processing to signal
Method solves the halfway technical problem of Signal separator while simplifying step, has reached the removal interference letter of very high degree of precision
Number effect.
The one side of the disclosure is to provide a kind of method that target jamming signal is removed from multiple signal, this method packet
It includes:
One group of input signal is received, every input signal in this group of input signal had both included useful signal or included interference letter
Number；
Improve the independence of the input signal；
Calculate the resulting coefficient matrix of independence for promoting the input signal；
Synchronize the input signal；
Input signal after synchronizing is separated into the channel containing target jamming signal and the channel without target jamming signal；
Intelligent selection is suitable, the channel without target jamming signal is exported as signal.
Another aspect of the present disclosure is to provide a kind of system that target jamming signal is removed from multiple signal, the system packet
It includes:
A set of input equipment for being used to input two-way or multiple signals；
Processor；And
The memory for storing computer-readable instruction, when the processor executes described instruction, which can be carried out:
Promote the independence of the input signal；
Calculate the resulting coefficient matrix of independence for promoting the input signal in input channel；
Synchronize the input signal；
Input signal after synchronizing is separated into the channel containing target jamming signal and the channel without target jamming signal；
Intelligent selection is suitable, the channel without target jamming signal is exported as signal.
On the other hand, the disclosure also provides a kind of non-transitory computer-readable storage medium, which is characterized in that the media storage
There is computer-readable instruction, when executed by a processor, it can be achieved that one kind removes target jamming signal from multiple signal
Method, which comprises
One group of input signal is received, every input signal in this group of input signal had both included useful signal or included interference letter
Number；
Promote the independence of the input signal；
Calculate the resulting coefficient matrix of independence for promoting the input signal；
Synchronize the input signal；
Input signal after synchronizing is separated into the channel containing target jamming signal and the channel without target jamming signal；
Intelligent selection is suitable, the channel without target jamming signal is exported as signal.
The application can eliminate or weaken asynchronization influence, improve signal source and extract performance, even if in useful signal and interference
It, also can be by removing interference signal persistently so as to improve the perceptibility of echo signal in the motion process of signal.
Detailed description of the invention
The description of example and not restrictive is carried out to embodiment of the present disclosure below with reference to accompanying drawings.Attached drawing is exemplary
And do not limited by the scale bar showed in figure.Identical or similar element uses identical symbol in different attached drawings
Label.
Fig. 1 is a kind of flow chart of method that target jamming signal is removed from multiple signal of the embodiment of the present disclosure；
Fig. 2 is the operating process drawing method one of the synchronization input signal of the embodiment of the present disclosure；
Fig. 3 is the operating process drawing method two of the synchronization input signal of the embodiment of the present disclosure；
Fig. 4 is the operating process drawing method three of the synchronization input signal of the embodiment of the present disclosure；
Fig. 5 is the operating process drawing method four of the synchronization input signal of the embodiment of the present disclosure；
Fig. 6 is a kind of Computer Systems Organization schematic diagram for realizing the embodiment of the present disclosure of the disclosure；
Fig. 7 is position view of the different sound sources to different sensors；
Fig. 8 shows the signal delay for the sensor that two have certain intervals.
Specific embodiment
Below in conjunction with the specific embodiment of the attached drawing detailed description disclosure.
Fig. 1 is a kind of process of method 1000 that target jamming signal is removed from input signal of the embodiment of the present disclosure
Figure.
In step 100, the signal that m signal source issues is received using n signal receiving device first, each signal connects
The received mixed signal of receiving apparatus is known as the input signal of the signal receiving device.Determine that every input signal (is m signal
Source issue signal mixed signal) in one or more signal sources issue signal be useful signal, other for interference letter
Number.The signal receiving device can be sensor or cloud platform.Signal receiving device is also possible to input data interface, and deposits
Storage unit is connected, and storage unit is previously stored with signal data, and input data interface receives signal data in storage unit.
In addition, every input signal can also include a variety of interference signals being not mutually identical.It is understood that in input signal
These interference signals be also possible to identical, the disclosure has no this specifically limited.For example, in the scene of electronic listening device
In, electronic listening device generally comprises at least two microphones, and each microphone can be used in receiving by sound generation source (useful letter
Number) and environmental background audio (interference signal) constitute mixed signal.It is useful since microphone is typically placed at different positions
Signal and interference signal are received by two or more microphones in mutually spaced different location, so by different Mikes
The environmental background audio that wind receives is discrepant each other in time domain and/or amplitude.For another example, film studio record and/
Or 360 in audio recording scene, audio is measured with two or more microphones, since microphone is typically placed at different positions
It sets, therefore useful signal and interference signal are received by two or more microphones in mutually spaced different location, by not
The environmental background sound that same microphone receives is discrepant each other in time domain and/or amplitude.For another example in brain-computer interface
In device scene, brain wave equipment generally comprises at least two electrodes, each electrode can receive comprising by E.E.G source signal and
The mixed signal that interference signal is constituted.Since electrode is typically placed at different positions, useful signal and interfering noise quilt
Two or more electrodes are received in mutually spaced different location, and the environmental noise received by different electrodes is in time domain
It and/or in amplitude is discrepant each other.Likewise, under water in echo detecting scene, echo reception device generally include to
Few two sensors, each sensor can be used in receiving the mixed signal from sound source and environmental noise.Since sensor is logical
Often be located differently, thus useful signal and interference signal by two or more sensors mutually it is spaced not
It is received with position, the environmental noise received by different sensors is discrepant each other in time domain and/or amplitude.Assuming that
There are two different sensor Mi, Mj and multiple and different signal source S1, S2 ... Sn, then signal received by Mi, Mj should be by
Following formula is constituted, and each different signal source travels to sensor Mi and Mj with different amplitude a and different time delays.
Mi=a1iS1(t1+τ1i)+a2iS2(t2+τ2i)+…+aniSn(tn+τni)
Mj=a1jS1(t1+τ1j)+a2jS2(t2+τ2j)+…+anjSn(tn+τnj)
Similarly, the signal that other sensors receive can also be analogized with same equation.
It is the position of two sensors and two signal sources in two-dimensional space as described in Figure 7 to simplify statement.It please infuse
Meaning, the figure are only to simplify to explain that all positions can be expanded to one-dimensional, three-dimensional or more high-dimensional table in stating in two-dimensional surface
It states.In order to more simplify statement, by taking acoustic signal as an example, it is assumed that S1, S2 are two sound sources, and M1, M2 are microphone.Assuming that sound
Spread speed is v, while assuming that the sample rate of the sensor is Fs.Therefore the propagation time of sound source to sensor can be used as follows
Formula indicates:
tij=Fs*dis { Si, Mj}/v (1)
In a random embodiment, v=34029cm/s, Fs=44.1kHZ.
Ideally, the energy of sound can be inversely proportional reduction with the increase of distance, then following equation can be used to indicate to pass
The voice signal that sensor receives:
Specific to described in Fig. 7, the formulae express is as follows, it is noted that in order to simplify statement, the formula is by all constant terms
It is reduced to 1.
Due in actual conditions, coefficient matrix shown in the right is unknown in S1, S2 and formula；The left one side of something of formula
M1realAnd M1realIt is mixed signal received by M1 and M2 microphone.In next step, step 200, decomposition coefficient matrix will mix
The a part for closing signal is reduced to useful signal.
In step 200, the independence of mixed signal is improved by decomposition coefficient matrix.Preferably, pass through resolving system
Matrix number maximizes the independence of mixed signal.The hypotheses of the present embodiment are as follows: each signal source is mutually indepedent
, then according to the statistical probability of central-limit theorem theory, (probability statistical distribution of the sum of i.e. multiple independent variables can be than every
The probability statistical distribution of one independent variable tends to normal distribution), judge the probability statistical distribution of embodiment mixed signal
Normal distribution can be tended to than the probability distribution of each signal source.Normal state in the present embodiment, passes through general as far as possible as a result,
The probability distribution statistical of mixed signal carrys out decomposition coefficient matrix far from normal distribution to improve the independence of signal source.Specifically,
Using coefficient matrix parameter as dependent variable, an objective function is set whether to calculate and measure variable close to normal distribution, is calculated
To the convergent optimal parameter of objective function to get resolution parameter matrix out.
Such as: step 200 choose lower array function as calculating and measure variable whether close to normal distribution objective function:
Kurt (y)=E { y4}-3(E{y2})2 (4)
E { } represents calculating desired value, and y is mixed signal.When target function value is 0, that is, show the probability distribution of y in just
State distribution.Kurtosis can certainly be substituted by other measurement modes as the standard far from normal distribution, the disclosure pair
There is no specific limitations for this.For this formula, objective function can be rewritten as following formula:
J(y)∝[E{G(y)}-E{G(v)}]2 (5)
Therefore using coefficient matrix parameter as dependent variable, above formula is objective function, is searched out by Newton iteration method
The convergent optimal parameter of objective function, i.e. resolution parameter matrix.Circular is briefly listed below:
1.Choose an initial(e.g.random)weight vector w.
2.Let w+=E { xg (wTx)}-E{g′(wTx)}w
3.Let w=w+/||w+||
4.If not converged, go back to 2.
Wherein g is the derived function of G.
Step 300, the input signal is synchronized in the time domain.The step can be realized by four kinds of different methods, in conjunction with attached
Fig. 2,3,4,5, this step 300 are described as follows.
As shown in Fig. 2, step 3101 is to intercept two or more discrete segments of interference signal, the duration of discrete segments
Control is at n milliseconds.If signal is audio signal, n needs to be greater than 0.98 millisecond, less than 20.03 milliseconds.When duration n is controlled at this
When in a section, the mankind are made to can't hear echo in the case where guaranteeing accuracy, therefore treatment effect is best in real time, user listens
Feel that effect is best.
Preferably, the discrete segments of each of step 3101 real-time continuous interception mixed signal.The method of this embodiment can
In real time to signal processing.
Then, it is directed to the mixed signal in each discrete segments section, the discrete patch is judged by way of pattern-recognition
Whether section is target jamming signal, and extracts target jamming signal.For example, being respectively male there are two sound source in acoustics case
People and woman, it is assumed that target jamming signal is male voice, then the pattern-recognition can be to the discrete segments of each n milliseconds of mixed signal
Whether be male voice, the snippet extraction is then carried out next step if male voice if making a decision, will judgement if interference signal is female voice
Next step is carried out for the snippet extraction of female voice.For another example two sound sources are respectively voice or non-voice.This field general technology
Personnel should be understood that other reasonable manners are also feasible.
The detection process of the interference signal of step 3101 can be by having detected interference signal from low level in n milliseconds
To high level, (that is, interference signal starts step signal response or from high level to low level, for example, setting man is sent out
Sound out is interference signal, does not need to finish entire word when man speaks, it is only necessary to detect the voice that this people speaks
The n millisecond of appearance determines that its sound is interference signal.This method is greatly reduced to complicated signal (such as sound letter
Number) requirement of detection process, thus reduce the complexity and its cost of calculating.
In step 3102, calculate two interference signal segments detected discrete-time convolution with obtain they when
Between postpone.Assuming that two mixed signals are respectively x, y, then the relevance formula between two signals is calculated are as follows:
Wherein, mx is the average value of x, and my is the average value of y, and d is time delay, and the molecular moiety of this formula is discrete time
Convolution.
Pass through different d, the i.e. selection of different time delay, relevance formula are as follows:
Based on this, that d chosen when maximum value generates in r (d) is time delay.
In step 3103, input signal described in the time delay d synchronization process based on acquisition.For example, if from
One input signal f1(t) the first interference signal for being detected in and from the second input signal f2(t) the second interference letter detected in
Number time delay be denoted as δ, then the first input signal f1(t) it is delayed by time δ, that is, is modified to f1(t- δ), it is thus defeated with second
Enter signal f2(t) synchronous.In another embodiment, if from the first input signal f1(t) interference signal that is detected in and from
Two input signal f2(t) time delay of the interference signal detected in is denoted as-δ, then the first input signal f1(t) it is synchronized to
f1(t+δ).Due to constantly real-time monitoring interference signal segment in this embodiment, the method can signal source and sensor not
When with mobile or relative movement, iteration time delay is ceaselessly updated, the variation of interference signal is dynamically tracked.
Referring to Fig. 3, in step 3201, since multiple sensors are placed in different positions, interference signal is by two
A or more sensor is received in mutually spaced different location.This embodiment is that calculate each interference signal in advance opposite
In the position of sensor, i.e., the relative delay of each interference signal；Then one is chosen according to the relative delay of each interference signal
A interference signal.Wherein, choosing an interference signal can also be by user's real-time selection.
Preferably, it is assumed that the distance between signal source to sensor 1 is d1, and arriving the distance between sensor 2 is d2, signal
Sample rate is Fs, signal velocity v.The calculation formula of its relative delay dir is as follows:
Dir=Fs* (d1-d2)/v (8)
Assuming that the distance between sensor is d, maximum direction Max (dir) is calculated by following formula:
Max (dir)=Fs*d/v (9)
If the above results are not integers, rounding-off method is taken to obtain integer.Then all directions have:-Max (dir) ..., -1,
0,1,…,Max(dir)。
Be detailed in the direction that distance in Fig. 8 describes, it is assumed that sample rate (Fs) is 48kHZ, two sensors (in the present example for
Acoustic signal, so this sensor is microphone) the distance between (d) be 2.47cm, the speed (v) that sound is propagated in air
For 340m/s, so maximum delay is 3.7 pieces of regions can be then divided into, respectively delay is -3, -2, -1,0,1,2,3.?
In Fig. 8 example, delay is then fixed as -3 by the region that such as default interference signal is -3 from delay.
Referring to Fig. 3, in step 3202, according to the interference signal region of user's real-time selection or preset interference signal area
Domain, extraction time delay.
Referring to Fig. 3, in step 3203, the time delay extracted according to 3202 does synchronization process, with 3103 steps.
Referring to fig. 4, this embodiment chooses to the interference signal from all relative delays.In step 3301, according to not
Same signal (such as sound), sensor distance, signal velocity carry out all time delays of analytical calculation.
Referring to fig. 4, in step 3302, all possible time delay τ is extracted1,τ2,…,τn,
Referring to fig. 4, in step 3303, each different time delay is repeated to do the synchronization in step 3103 respectively
Change processing.
Referring to Fig. 5, in step 3401, pass through user's real-time selection or preset useful signal direction.
Referring to Fig. 5, in step 3402, the time delay in these directions is calculated.
Referring to Fig. 5, in the method based on the obtained all senses of Fig. 4, in step 3403, by these useful signals
Time delay excluded in all possible direction, each remaining different time delay is repeated to do step respectively
Synchronization process in 3103.Referring again to Fig. 1, in step 400, the input signal after synchronizing is separated into containing target jamming
The channel of signal and channel without target jamming signal.Preferably, step 400 is by synchronized signal matrix and step
The multiplyings of 200 coefficient matrixes determined is realized.
For example, referring to the example of step 100, it is assumed that mixed signal composition is as follows:
After the coefficient matrix obtained by step 200, it is multiplied with the signal matrix after synchronization, formula is as follows:
By the formula it is found that two channels, one of channel will be generated
In other words, the channel is by the S1 of the S2 comprising 96% and 4%, and if target jamming signal is S1, which should
It is selected and exports.Therefore, in this example, synchronized separating effect has reached 96%.
How to select, can be unfolded in this two channels in step 500.
Similarly, if if target jamming signal is S2, the mixed signal of synchronous S2 is multiplied with coefficient matrix.It selects simultaneously
Select suitable channel output.
It, in step 500, can be according to different signal phases based on two channels obtained in step 400 referring to Fig. 1
Energy is selected, that relatively low channel of selection signal energy is output channel.The calculation method of signal energy can be with
It is the root-mean-square value of the signal.The selection course apply to the channel containing target jamming signal obtained in the step 500 and
In channel without target jamming signal.
Further, in the embodiment of Figure 4 and 5, it will an output channel is generated in delay in different times,
In the embodiment of Fig. 4, optimal channel is detected and selected as signal output (as generated the target jamming signal in channel based on feature
The least channel of ingredient)；Optimal channel can be selected as signal output (as produced according to signal energy in the 5 embodiment of figure 5
The minimum channel of target jamming signal energy in raw channel).
Preferably, after step 500 has separated interference signal, can also include to after separation useful signal and interference
The step of signal is further processed, for example do frequency domain enhancing.For example, in hearing aid application, it can be by the useful sound after separation
Frequency signal does personalized frequency domain enhancing.
In one embodiment, the disclosure provides a kind of device comprising processor and man-machine interactively interface.The device
It can also include but is not limited to memory, controller, input/output module, information receiving module.Processor is for executing
State step 100,200,3201-3203 (or 3401-3403) and 400,500 and frequency domain enhancing (optional).User passes through
Man-machine interactively interface selects him to wish that region for interference signal region in real time.Man-machine interactively interface includes but are not limited to
Speech reception module, sensor, video reception module.Touch screen, keyboard, button, knob project interface, the virtual interface 3D.With
Family includes by phonetic order by way of the real-time selection of man-machine interactively interface, and different gestures or movement by user lead to
Cross the region of selection different identification.When man-machine interactively interface is touch screen, user can click wherein some region, the disclosure and mention
For a kind of machine of user controllable selectable removal interference signal, and delay can be adjusted in real time.
Above-mentioned steps 100-400 may occur to be different from sequencing described in attached drawing.For example, step 100 and step
The sequence of rapid 300 second embodiment (i.e. 3201-3203) can exchange mutually.For another example in practical applications, according to tool
The function that body is related to, any two step in step 100-400 parallel execution or may execute in reverse order.
Preferably, step 200 is implemented before step 300, i.e., then first design factor matrix synchronizes input in the time domain
Signal.Doing so is advantageous in that, does not need all to recalculate coefficient of first order matrix according to different time delays.So may be used
Largely to save calculation amount.Especially in Fig. 4 and Fig. 5 described embodiment, it is only necessary to calculate coefficient of first order matrix, so that it may
To obtain result.Meanwhile the disclosure concludes that the mixed signal after synchronizing is calculated by carrying out a large amount of experiment
Coefficient matrix and the calculated coefficient matrix of original mixed signal are very nearly the same.Therefore, this method is in not loss coefficient matrix
Precision while substantial saved calculation amount.
Preferably, in step 100, after receiving n input signal using m signal receiving device, according to judgement
Whether conditional decision removes the received input signal of one or more signal receiving device institutes in multiple signal receiving devices.
In some embodiment, input signal is acoustic signal, signal receiving device be acoustic signal reception device (such as
Microphone).(wherein L is the discrete signal length of interception, and X is any two acoustics letter when Rule of judgment is Fs*X/V < L/3
The distance of number reception device, V is signal velocity, and Fs is sample rate), by wherein one in two acoustic signal reception devices
A received acoustic signal removal.The present embodiment reduces while the accuracy that not Effect Mode identifies to be needed to calculate
Data volume, improve computational efficiency, reduce power consumption.
The signal includes audio signal, picture signal, electromagnetic signal, eeg signal, electric signal, radio wave signal
And other forms can be by sensor received signal, the disclosure has no specific limitation to this.
The disclosure can greatly promote the perceptibility of echo signal while reduce operation cost.In addition, disclosure input letter
Number it have passed through synchronization process in the time domain, therefore disclosed method minimizes frequency distortion.
Fig. 6 is a kind of structural schematic diagram of computer system 3000 for being adapted to carry out above embodiment of the disclosure.
As shown in fig. 6, the computer system 3000 includes central processing unit (CPU) 3001, which can be according to storage
Program instruction in electric programmable read-only memory (EPROM) 3002 or random access memory (RAM) 3003 executes various
Suitable operation and process.The necessary journey for running the system 3000 can also be stored on random access memory (RAM) 3003
Sequence and data.The central processing unit (CPU) 3001, electric programmable read-only memory (EPROM) 3002 and random access memory
Device (RAM) 3003 is connected with each other by bus 3004.The bus 3004 is also connected with input/output (I/O) interface 3005.Institute
It states bus 3004 and is also connected with direct memory access (DMA) interface 3004 to accelerate data exchange.
Input/output (I/O) interface 3005 is also connected with following elements: movable data memory 3007, including
USB storage, solid state hard disk etc.；Wireless data transfer line road 3008, including local area network (LAN), bluetooth, near-field communication equipment
(NFC)；And the signal converter 3009 being connected with data input channel 2010 and data output channel 3011.According to this public affairs
Another embodiment opened, the process that above-mentioned flow chart is related to can be by similar with the computer system 3000 without keyboard, mouse
The embedded computer system of mark and hard disk is realized.Wireless data transfer line road 3008 or movable data memory 3007 are advantageous
Upgrade in the update of program.
Wherein, above-mentioned processor can be cloud processor, and memory can be cloud memory.
Further, according to the another embodiment of the disclosure, the process that above-mentioned flow chart is related to can be by computer software journey
Sequence is realized.For example, the embodiment of the present disclosure provides a kind of computer program product comprising be stored in tangible machine readable medium
Computer program, the program include the program code of method shown in execution flow chart.In the present embodiment, the computer journey
Sequence can be downloaded and be installed by wireless data transfer line road 3008, and/or be installed from removable medium 3007.
Flow chart in attached drawing is produced with system, method and computer program in block diagrams disclosure difference embodiment
Framework, function and the realization of the operation process of product.In this regard, each of flow chart and block diagram box represent a kind of mould
Block, program segment or code unit.The module, section or code unit includes one or more for realizing specified logic
The executable instruction of function.It should be pointed out that the function that module indicates may be attached to be different from certain preferred embodiments
Sequencing described in figure occurs.For example, in practical applications, according to the function being specifically related to, two connected box institutes
The possible parallel execution of the operation shown executes in reverse order.It should be noted simultaneously that in flowchart and/or the block diagram
Each box or combinations thereof can pass through dedicated, based on hardware a, executable specific function or the system of operation
It realizes, or is realized by the combination of special-purpose software and computer instruction.
Unit involved in the embodiment of the present disclosure or module can be pacified by software or hardware realization, the unit or module
Dress is in the processor.The title of the unit or module should not cause to limit to itself of unit or module.
On the other hand, the disclosure also provides a kind of computer readable storage medium.The computer readable storage medium can
The computer readable storage medium in instrument and equipment to be mounted in above-described embodiment application, is also possible to independent not fill
Assigned in the computer readable storage medium in instrument and equipment.The computer-readable recording medium storage one or more program,
Described program is performed by one or more processors the method that echo signal is sorted in the slave noise signal to realize the disclosure.
The above content is combine specific preferred embodiment to made by the disclosure further description, and it cannot be said that
The specific implementation of the disclosure is only limited to these instructions.For disclosure person of an ordinary skill in the technical field,
Under the premise of not departing from disclosure design, a number of simple deductions or replacements can also be made, all shall be regarded as belonging to the disclosure
Protection scope.
Claims (16)
1. a kind of method for removing target jamming signal from multiple signal, which is characterized in that this method comprises:
One group of input signal is received, every input signal in this group of input signal had both included useful signal or included interference letter
Number；
Improve the independence of the input signal；
Calculate the resulting coefficient matrix of independence for promoting the input signal；
Synchronize the input signal；
Input signal after synchronizing is separated into the channel containing target jamming signal and the channel without target jamming signal；
Channel of the intelligent selection without target jamming signal is exported as signal.
2. the method according to claim 1, wherein the operation for synchronizing the input signal includes:
Detect the interference signal segment in every input signal；
Discrete-time convolution operation is carried out to obtain relative time-delay to the every two interference signal segment detected；
Based on acquired time delay, the input signal is synchronized；
Selection is marked as the signal priority direction of interference signal；
Calculate the relative time-delay of the interference signal from privileged direction；
Based on preset time delay, the input signal is synchronized；
Selection is marked as all possible sense of interference signal；
A series of time delays are estimated, τ 1, τ 2 ..., τ n are denoted as；
Based on a series of time delay, the input signal is synchronized；
Selection is marked as the signal approach axis of useful signal；
Determine the time delay of the interference signal from remaining direction；
Based on determining time delay, the interference signal is synchronized.
3. the method according to claim 1, wherein the sustainable upgrading of the synchronization of the input signal is to fit
The motion state in induction signal source.
4. the method according to claim 1, wherein the input signal is taken from the site being spaced apart from each other.
5. the method according to claim 1, wherein the independence for improving the input signal includes: logical
Cross the Gaussian Profile that independent component analysis maximizes the input signal.
6. according to the method described in claim 2, it is characterized in that, the interference signal segment detected in every input signal
It include: the interference signal segment detected by pattern-recognition in every input signal.
7. the method according to claim 1, wherein the input signal is by sensor received signal.
8. the method according to the description of claim 7 is characterized in that the input signal is one of following:
Audio signal；
Electric signal；
Picture signal；And
Radiofrequency signal.
9. a kind of system for removing target noise from signal, characterized in that it comprises:
A set of input equipment for being used to input one group of input signal；
Processor；And
The memory for storing computer-readable instruction, when the processor executes described instruction, which can be carried out:
Promote the independence of the input signal；
Calculate the resulting coefficient matrix of independence for promoting the input signal in input channel；
Synchronize the input signal；
Input signal after synchronizing is separated into the channel containing target jamming signal and the channel without target jamming signal；
Channel of the intelligent selection without target jamming signal is exported as signal.
10. system according to claim 9, which is characterized in that described to synchronize the input signal and include:
Detect the interference signal segment in every input signal；
Discrete-time convolution operation is carried out to obtain relative time-delay to the every two interference signal segment detected；
Based on acquired time delay, the input signal is synchronized；
Selection is marked as the signal priority direction of interference signal；
Calculate the relative time-delay of the interference signal from privileged direction；
Based on preset time delay, the input signal is synchronized；
Selection is marked as all possible sense of interference signal；
A series of time delays are estimated, τ 1, τ 2 ..., τ n are denoted as；
Based on a series of time delay, the input signal is synchronized；
Selection is marked as the signal approach axis of useful signal；
Determine the time delay of the interference signal from remaining direction；
Based on determining time delay, the interference signal is synchronized.
11. system according to claim 9, which is characterized in that the input signal is taken from the site being spaced apart from each other.
12. system according to claim 9, which is characterized in that the independence for promoting the input signal includes: logical
Cross the Gaussian Profile that independent component analysis maximizes the input signal.
13. system according to claim 10, which is characterized in that the interference signal piece in every input signal of the detection
Section includes: the interference signal segment detected in every input signal by pattern-recognition.
14. system according to claim 9, which is characterized in that the input signal is by sensor received signal.
15. system according to claim 14, which is characterized in that the input signal is one of following:
Audio signal；
Electric signal；
Picture signal；
Radiofrequency signal.
16. a kind of non-transitorycomputer readable storage medium, which is characterized in that the storage medium is stored with instruction, works as processing
, it can be achieved that a kind of method for separating target jamming signal from multiple signal when device executes the instruction, which comprises
One group of input signal (observation signal) is received, every input signal contains target jamming signal；
Promote the independence of the input signal；
Calculate the resulting coefficient matrix of independence for promoting the input signal；
Synchronize the input signal；
Input signal after synchronizing is separated into the channel containing target jamming signal and the channel without target jamming signal；
Channel of the intelligent selection without target jamming signal is exported as signal.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710698651.6A CN109413543B (en) | 2017-08-15 | 2017-08-15 | Source signal extraction method, system and storage medium |
| EP17921701.3A EP3672275A4 (en) | 2017-08-15 | 2017-12-21 | Method and system for extracting source signal, and storage medium |
| PCT/CN2017/117813 WO2019033671A1 (en) | 2017-08-15 | 2017-12-21 | Method and system for extracting source signal, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710698651.6A CN109413543B (en) | 2017-08-15 | 2017-08-15 | Source signal extraction method, system and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109413543A true CN109413543A (en) | 2019-03-01 |
| CN109413543B CN109413543B (en) | 2021-01-19 |
Family
ID=65362112
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710698651.6A Active CN109413543B (en) | 2017-08-15 | 2017-08-15 | Source signal extraction method, system and storage medium |
Country Status (3)
| Country | Link |
|---|---|
| EP (1) | EP3672275A4 (en) |
| CN (1) | CN109413543B (en) |
| WO (1) | WO2019033671A1 (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2430975A1 (en) * | 2010-09-17 | 2012-03-21 | Stichting IMEC Nederland | Principal component analysis or independent component analysis applied to ambulatory electrocardiogram signals |
| CN102750952A (en) * | 2011-04-18 | 2012-10-24 | 索尼公司 | Sound signal processing device, method, and program |
| CN103083012A (en) * | 2012-12-24 | 2013-05-08 | 太原理工大学 | Atrial fibrillation signal extraction method based on blind source separation |
| CN103426435A (en) * | 2012-05-04 | 2013-12-04 | 索尼电脑娱乐公司 | Source separation by independent component analysis with moving constraint |
| CN103426434A (en) * | 2012-05-04 | 2013-12-04 | 索尼电脑娱乐公司 | Source separation by independent component analysis in conjunction with source direction information |
| CN103654840A (en) * | 2012-08-29 | 2014-03-26 | 索尼公司 | Signal processing system, signal processing apparatus, and storage medium |
| CN104053107A (en) * | 2014-06-06 | 2014-09-17 | 重庆大学 | Hearing aid device and method for separating and positioning sound sources in noise environments |
| CN104091356A (en) * | 2014-07-04 | 2014-10-08 | 南京邮电大学 | X-ray medical image objective reconstruction based on independent component analysis |
| CN105640500A (en) * | 2015-12-21 | 2016-06-08 | 安徽大学 | Scanning signal feature extraction method based on independent component analysis and recognition method |
| CN105996993A (en) * | 2016-04-29 | 2016-10-12 | 南京理工大学 | System and method for intelligent video monitoring of vital signs |
| CN106356075A (en) * | 2016-09-29 | 2017-01-25 | 合肥华凌股份有限公司 | Blind voice separation method and structure, voice control system and electric appliance assembly |
| WO2017084397A1 (en) * | 2015-11-19 | 2017-05-26 | The Hong Kong University Of Science And Technology | Method, system and storage medium for signal separation |
| CN107025446A (en) * | 2017-04-12 | 2017-08-08 | 北京信息科技大学 | A kind of vibration signal combines noise-reduction method |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100372277C (en) * | 2006-02-20 | 2008-02-27 | 东南大学 | Space time separation soft inputting and outputting detecting method based on spatial domain prewhitening mergence |
| CN100495388C (en) * | 2006-10-10 | 2009-06-03 | 深圳市理邦精密仪器有限公司 | Signal processing method using space coordinates convert for realizing signal separation |
| US9100734B2 (en) * | 2010-10-22 | 2015-08-04 | Qualcomm Incorporated | Systems, methods, apparatus, and computer-readable media for far-field multi-source tracking and separation |
| CN102571296B (en) * | 2010-12-07 | 2014-09-03 | 华为技术有限公司 | Precoding method and device |
| CN102868433B (en) * | 2012-09-10 | 2015-04-08 | 西安电子科技大学 | Signal transmission method based on antenna selection in multiple-input multiple-output Y channel |
| CN103197183B (en) * | 2013-01-11 | 2015-08-19 | 北京航空航天大学 | A kind of method revising Independent component analysis uncertainty in electromagnetic interference (EMI) separation |
-
2017
- 2017-08-15 CN CN201710698651.6A patent/CN109413543B/en active Active
- 2017-12-21 WO PCT/CN2017/117813 patent/WO2019033671A1/en unknown
- 2017-12-21 EP EP17921701.3A patent/EP3672275A4/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2430975A1 (en) * | 2010-09-17 | 2012-03-21 | Stichting IMEC Nederland | Principal component analysis or independent component analysis applied to ambulatory electrocardiogram signals |
| CN102750952A (en) * | 2011-04-18 | 2012-10-24 | 索尼公司 | Sound signal processing device, method, and program |
| CN103426435A (en) * | 2012-05-04 | 2013-12-04 | 索尼电脑娱乐公司 | Source separation by independent component analysis with moving constraint |
| CN103426434A (en) * | 2012-05-04 | 2013-12-04 | 索尼电脑娱乐公司 | Source separation by independent component analysis in conjunction with source direction information |
| CN103654840A (en) * | 2012-08-29 | 2014-03-26 | 索尼公司 | Signal processing system, signal processing apparatus, and storage medium |
| CN103083012A (en) * | 2012-12-24 | 2013-05-08 | 太原理工大学 | Atrial fibrillation signal extraction method based on blind source separation |
| CN104053107A (en) * | 2014-06-06 | 2014-09-17 | 重庆大学 | Hearing aid device and method for separating and positioning sound sources in noise environments |
| CN104091356A (en) * | 2014-07-04 | 2014-10-08 | 南京邮电大学 | X-ray medical image objective reconstruction based on independent component analysis |
| WO2017084397A1 (en) * | 2015-11-19 | 2017-05-26 | The Hong Kong University Of Science And Technology | Method, system and storage medium for signal separation |
| CN105640500A (en) * | 2015-12-21 | 2016-06-08 | 安徽大学 | Scanning signal feature extraction method based on independent component analysis and recognition method |
| CN105996993A (en) * | 2016-04-29 | 2016-10-12 | 南京理工大学 | System and method for intelligent video monitoring of vital signs |
| CN106356075A (en) * | 2016-09-29 | 2017-01-25 | 合肥华凌股份有限公司 | Blind voice separation method and structure, voice control system and electric appliance assembly |
| CN107025446A (en) * | 2017-04-12 | 2017-08-08 | 北京信息科技大学 | A kind of vibration signal combines noise-reduction method |
Also Published As
| Publication number | Publication date |
|---|---|
| EP3672275A1 (en) | 2020-06-24 |
| CN109413543B (en) | 2021-01-19 |
| EP3672275A4 (en) | 2023-08-23 |
| WO2019033671A1 (en) | 2019-02-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Deep learning based binaural speech separation in reverberant environments | |
| US10063965B2 (en) | Sound source estimation using neural networks | |
| Hoshen et al. | Speech acoustic modeling from raw multichannel waveforms | |
| Ono et al. | Blind alignment of asynchronously recorded signals for distributed microphone array | |
| Katz et al. | A comparative study of interaural time delay estimation methods | |
| Mandel et al. | An EM algorithm for localizing multiple sound sources in reverberant environments | |
| EP2891955A1 (en) | In-vehicle gesture interactive spatial audio system | |
| CN111063342B (en) | Speech recognition method, speech recognition device, computer equipment and storage medium | |
| Ma et al. | Exploiting deep neural networks and head movements for binaural localisation of multiple speakers in reverberant conditions | |
| JP2014215461A (en) | Speech processing device, method, and program | |
| Nguyen et al. | SALSA-Lite: A fast and effective feature for polyphonic sound event localization and detection with microphone arrays | |
| Kürby et al. | Bag-of-Features Acoustic Event Detection for Sensor Networks. | |
| JP7214798B2 (en) | AUDIO SIGNAL PROCESSING METHOD, AUDIO SIGNAL PROCESSING DEVICE, ELECTRONIC DEVICE, AND STORAGE MEDIUM | |
| Pujol et al. | BeamLearning: An end-to-end deep learning approach for the angular localization of sound sources using raw multichannel acoustic pressure data | |
| Jiang et al. | Acoustic scene classification using ensembles of convolutional neural networks and spectrogram decompositions | |
| CN112180318A (en) | Sound source direction-of-arrival estimation model training and sound source direction-of-arrival estimation method | |
| CN109413543A (en) | A kind of source extraction method, system and storage medium | |
| Cobos et al. | Two-microphone separation of speech mixtures based on interclass variance maximization | |
| Jafari et al. | Underdetermined blind source separation with fuzzy clustering for arbitrarily arranged sensors | |
| CN110441730A (en) | A kind of microphone array sound source direction system based on analog signal processing framework | |
| Bouafif et al. | TDOA Estimation for Multiple Speakers in Underdetermined Case. | |
| Masnadi-Shirazi et al. | Separation and tracking of multiple speakers in a reverberant environment using a multiple model particle filter glimpsing method | |
| Deleforge et al. | Audio-motor integration for robot audition | |
| Oualil et al. | Joint detection and localization of multiple speakers using a probabilistic interpretation of the steered response power | |
| CN111429937A (en) | Voice separation method, model training method and electronic equipment |
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 |