CN110051325A - Electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD - Google Patents
Electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD Download PDFInfo
- Publication number
- CN110051325A CN110051325A CN201910246351.3A CN201910246351A CN110051325A CN 110051325 A CN110051325 A CN 110051325A CN 201910246351 A CN201910246351 A CN 201910246351A CN 110051325 A CN110051325 A CN 110051325A
- Authority
- CN
- China
- Prior art keywords
- signal
- noise
- eemd
- frequency
- electrocardiosignal
- 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.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7253—Details of waveform analysis characterised by using transforms
- A61B5/726—Details of waveform analysis characterised by using transforms using Wavelet transforms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
- G06F2218/06—Denoising by applying a scale-space analysis, e.g. using wavelet analysis
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Cardiology (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
A kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD is claimed in the present invention.For noise behavior in electrocardio, Main Noise Sources are divided into low-frequency disturbance and High-frequency Interference.High-frequency Interference includes power frequency and myoelectricity interference, and design wavelet thresholding method realizes the elimination of High-frequency Interference；Low-frequency disturbance, that is, baseline drift interference provides the foundation that the size of important parameter addition aid in noise and set average time determine in EEMD algorithm, then eliminates baseline drift using improvement EEMD algorithm and interfere for the deficiency of EEMD algorithm.The case where existing simultaneously for interference a variety of in electrocardiosignal devises the electrocardiosignal based on wavelet transformation and improvement EEMD and disposably integrates denoising algorithm.The present invention can be improved signal-to-noise ratio, reduce mean square deviation, keep ecg wave form feature.
Description
Technical field
The invention belongs to vital sign parameter signals processing technology fields, and specifically a kind of electrocardiosignal integrates de-noising side
Method.
Background technique
One direct join of cardiovascular disease threatens the major disease of human life and health by persistently becoming.According to World Health Organization's number
According to display, the people in worldwide every year there are about 20,000,000 dies of cardiovascular disease.In China, with adding for aging of population
Speed, cardiovascular morbidity number is by sustainable growth.Cardiovascular disease has become China or even global public health problem, it
As the number one killer of human health, government and the extensive attention of the public have been caused.Electrocardiosignal is as characterization human body disease
The bioelectrical signals for managing information, are one of most important physiology signs of human body, it is able to reflect body parts especially
The operating status of heart, prevention and diagnosis for cardiovascular disease all have greatly value.
ECG Signal Analysis is one and is related to the very wide project in face, includes pretreatment, feature extraction, heart rate variability
Analysis etc. research contents.On the basis that the present invention is studied based on forefathers, carries out in terms of ECG signal processing and grind
Study carefully work, it is intended to provide theoretical direction for the clinical application and development of new ecg analysis system of ECG Signal Analysis.
Electrocardio de-noising is the preprocessing process of ecg analysis.Electrocardiosignal is a kind of very representational biomedical letter
Number, it is during acquisition, it will usually have many noise couplings into original signal, processing and analysis band to electrocardiosignal
Come greatly difficult.Therefore the electrocardiosignal that " clean " how is extracted from signals and associated noises, guarantees the accuracy one of medical diagnosis
It is directly the vital task of ECG signal processing.Wherein mainly there are three big noise sources: myoelectricity, power frequency and needle position misalignment interference.
In all kinds of interference, influence of the baseline drift to electrocardiosignal is maximum.As a kind of main noise source, it is a kind of low frequency letter
Number, it is generally the case that its frequency is less than 1Hz.For electrocardiosignal, its own is also containing the low-frequency component of very abundant, base drift
Interference is superimposed with it can cover useful information therein, to the characteristic wave detection of subsequent electrocardiosignal and heart rate variability analysis
It produces serious influence.Currently, most of ECG signal processings are concentrated mainly on the elimination for individually interfering, for electrocardio
The integrated filter technique study of signal be not also it is very deep, the present invention wishes that a variety of noise jammings can be removed simultaneously by finding out one kind
Denoising of ECG Signal.This method can not only preferably remove a variety of noise jammings, and guarantee the feature of electrocardiosignal
Waveform provides good basis for the analysis of subsequent signal.
For the processing of electrocardiosignal, many outstanding analysis methods are had been proposed in experts and scholars, are occurred a large amount of
Research achievement.They are mainly analyzed and processed signal from time domain, frequency domain and the several angles of time-frequency domain, these research sides
Method provides great power-assisted for life the reach of science.However, existing ECG Signal Analysis method there are still effects not
Good problem, there are also very big rooms for improvement in theoretical research and clinical application.This just needs more in-depth study.This
Mainly using several important signals such as wavelet transformation and empirical mode decomposition researchs and analyses method for invention, they are all novel letters
Number analysis method has been widely used in many fields, and to their research and application also in progress like a raging fire.
Wavelet transformation is a kind of Time-Frequency Analysis Method, the analysis suitable for non-stationary signal.Currently, it has been used in respectively
A field, especially in terms of signal and image denoising.For the non-stationary property of electrocardiosignal, wavelet transformation becomes one kind very
Effective analysis means.Although wavelet transformation has been achieved with many achievements in terms of electrocardio processing, about it research also
It is far from terminating.Wavelet transformation is mainly used for removing the High-frequency Interference in electrocardio in the present invention.
In recent years, it is engaged in the researcher of ECG's data compression method, energy is mainly concentrated on into Huang et al. in 1998
Empirical mode decomposition (Empirical Mode Decomposition, EMD) algorithm that year proposes, it can decompose sophisticated signal
For the sum of one group of Intrinsic mode function (Intrinsic Mode Function, IMF), however after this algorithm is to signal decomposition
It is frequently present of modal overlap problem.
In order to overcome the problems, such as that existing modal overlap after EMD algorithm decomposes, Z.H.Wu etc. propose set empirical modal point
(Ensemble Empirical Mode Decomposition, the EEMD) method of solution, the basic principle of this method is in original letter
Number upper be added assists white noise several times, using noise mixture as a signal to be decomposed, when original signal is added uniformly point
When the white noise background of cloth, the signal of different time scales, which will be mapped to, suitably to be referred on scale, then again respectively to the letter
Mixture of making an uproar carries out EMD resolution process, the true mode that last averaged can be approached.It will be improved in the present invention
EEMD algorithm is for removing baseline drift interference.
In conclusion the Main Noise Sources in electrocardio are divided into High-frequency Interference and low-frequency disturbance by the present invention.High-frequency Interference packet
Power frequency and myoelectricity interference are included, design wavelet thresholding method realizes the elimination of High-frequency Interference；Low-frequency disturbance, that is, baseline drift interference, for
The deficiency of EEMD algorithm, provide important parameter in EEMD algorithm the size of aid in noise is added and set average time determines according to
According to then using the elimination baseline drift interference of improved EEMD algorithm.It is finally directed to a variety of interference in electrocardiosignal while depositing
The case where, the present invention devises the filter algorithm that can disposably eliminate a variety of noises.
Currently, show traditional analysis method in terms of ECG's data compression method and continue to bring extra power into play, new signal
The phenomenon that analysis method continues to bring out.With the progress of science and technology and going deep into for research, it is believed that in the near future can be real
The high efficiency smart processing of existing signal.
Summary of the invention
Present invention seek to address that the above problem of the prior art.Propose a kind of solution electrocardiosignal traditional filtering method effect
Fruit is bad and wave distortion based on wavelet transformation and improve the electrocardiosignal integrated filter method of EEMD.Technical side of the invention
Case is as follows:
A kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD comprising following steps:
Step 1: establishing noisy electrocardiosignal model: 103 electrocardiosignal of MIT-BIH database is chosen, wherein 103 electrocardios
Signal is influenced by baseline drift than more serious, and the interference including power frequency, myoelectricity, baseline drift is added, and obtains noisy electrocardio
Signal x'；
Step 2: wavelet threshold eliminates high-frequency noise: noisy electrocardiosignal x' is decomposed using wavelet basis coif3,
High-frequency noise is eliminated using threshold method, the threshold method uses soft-threshold mode, and selection of threshold function " Rigrsure " is eliminated high
Signal sequence after frequency noise is denoted as x；
Step 3: use experience mode decomposition (EMD) decomposes the signal x after eliminating high-frequency noise, by first
Intrinsic mode function component obtains the value for improving parameter k and N in EEMD algorithm as high fdrequency component；Wherein k be white noise with
The ratio of signal amplitude standard deviation, N are ensemble average number；
Step 4: being decomposed using EEMD algorithm is improved: after determining parameter, using improved EEMD algorithm to elimination
Signal sequence x after high-frequency noise is decomposed；Improved EEMD provides important parameter in EEMD algorithm and aid in noise is added
Size and set average time determine foundation, the parameter k and N of original EEMD entirely by be manually set, none determine
Foundation, according to discriminate | mean (IMFi) | > ε chooses the IMF component for the condition that meets as the component part for extracting base drift, obtains
To base drift signal be denoted as；
Step 5: obtaining the baseline drift signal extracted: after signal decomposition, asking equal to all IMF components and residual volume
Value, according to discriminate | mean (IMFi) | > ε chooses the IMF component for the condition that meets as the component part for extracting base drift, obtains
Base drift signal be denoted as f；
Step 6: reconstruct obtains the electrocardiosignal after de-noising: subtracting the base drift letter of extraction using the signal x after elimination high frequency
Number f, the signal y=x-f after de-noising can be obtained.
Further, the step 1 chooses pure electrocardiosignal " MIT-BIH Noise Stress Test
The sample rate of 103 signals in Database " database, MLII lead, signal is 360Hz, takes 4096 data points.
Further, the interference including power frequency, myoelectricity, baseline drift is added in the step 1, specifically includes: simulation power frequency
Interference: power frequency interference signals are 0.12mV using amplitude, and frequency is the sinusoidal signal of 50Hz；Simulate myoelectricity interference: myoelectricity interference
Use white noise；
Baseline drift interference: the baseline drift interference of introducing has the drift of simulation base and true base to float two kinds of forms, simulation base drift
Use amplitude for 0.2mV, frequency is the sinusoidal signal of 0.5Hz, and true base, which winnows, takes " MIT-BIH Noise Stress Test
Base in Database " database floats interference signal.
Further, in the step 2, mainly include three kinds of situations in the threshold value quantizing step of wavelet transformation: (1) writing from memory
Recognize threshold value；(2) given threshold value；(3) forcing denoising is wavelet coefficient zero.
Further, in the step 3, the value for improving parameter k and N in EEMD algorithm is obtained, is specifically included: be added
White noise must meet following condition:
In formula, the radio-frequency component standard deviation of A --- signal and the proportionality coefficient of original signal standard deviation；It takes
Ensemble average times N and proportionality coefficient k should meet following relationship:
First IMF component that selection signal obtains after the decomposition of EMD method is as signal radio-frequency component, to obtain
Coefficient A；Then coefficient k is obtained according to formula (1)；Relative error e is preset simultaneously according to formula (2) available ensemble average
Number N.
Further, the step 4 is decomposed using EEMD algorithm is improved: after determining parameter, use is improved
EEMD algorithm decomposes the signal sequence x after eliminating high-frequency noise；Specific step is as follows:
(1) auxiliary white noise signal+k σ is added in the signal x of removal high-frequency noisexN (t), wherein k is white noise
The ratio of sound and signal amplitude standard deviation, σxPoor for signal standards, n (t) is normalization white noise, constitutes noise mixture:
X1(t)=x (t)+k σx·n(t) (3)
(2) EMD decomposition is carried out to noise mixture:
Wherein cjIndicate j-th of intrinsic mode function, r that noise mixture generates after EMD is decomposedmIndicate noise mixing
Trend term of the body after EMD is decomposed, m indicate the quantity of intrinsic mode function after EMD decomposition；
(3) step (1) and (2) is repeated, different white noises is added every time:
Xi(t)=x (t)+kni(t) (5)
Resolve into IMF:
ci,jIndicate that the noise mixture after assisting white noise is added through EMD in i-th in the signal x of removal high-frequency noise
J-th of intrinsic mode function, r after decompositioni,mAfter indicating that auxiliary white noise is added in i-th in the signal x of removal high-frequency noise
Trend term of the noise mixture after EMD is decomposed；
(4) n times are repeated, are then averaged to each IMF:
(5) last decomposition result are as follows:
It advantages of the present invention and has the beneficial effect that:
Beneficial effect of the present invention essentially consists in and improves signal-to-noise ratio, reduces mean square deviation and keep two sides of ecg wave form feature
Face.It is specific as follows:
1. improving signal-to-noise ratio, reducing mean square deviation
Electrocardiosignal is a kind of very representational biomedicine signals, it is during acquisition, it will usually have very
For more noise couplings into original signal, processing and analysis to electrocardiosignal bring great difficulty.It is filtered for conventional method
Ineffective problem, the present invention are applied to the elimination of electrocardio noise using novel signal processing method, are filtered using wavelet transformation
Except high-frequency noise, baseline drift interference, relative to other methods, method energy proposed by the present invention are eliminated using EEMD algorithm is improved
Signal-to-noise ratio is enough improved, mean square deviation is reduced, denoising effect is more preferable.
2. keeping ecg wave form feature
In all kinds of interference, influence of the baseline drift to electrocardiosignal is maximum.As a kind of main noise source, it is one
Kind low frequency signal, it is generally the case that its frequency is less than 1Hz.For electrocardiosignal, its own also the low frequency containing very abundant at
Point, base drift interference is superimposed with it can cover useful information therein, become to the characteristic wave detection of subsequent electrocardiosignal and heart rate
Specific analysis produces serious influence.The present invention carries out the elimination of baseline drift interference, this method energy using EEMD algorithm is improved
The loss for enough reducing low-frequency component, keeps electrocardiosignal signature waveform.
Detailed description of the invention
Fig. 1 is that the present invention provides the comprehensive noise-eliminating method the general frame of electrocardiosignal of preferred embodiment.
For 103 signals in " MIT-BIH Arrhythmia Database " database, (MLII lead takes 4096 to Fig. 2
Data point)；
Fig. 3 is noisy electrocardiosignal model: (a) 103 signals (b) are superimposed Hz noise (c) and are superimposed myoelectricity interference (d) superposition
It simulates true base drift (f) superposed simulation base drift of base drift (e) superposition, power frequency and myoelectricity (g) and is superimposed true base drift, power frequency and myoelectricity；
Fig. 4 is the effect picture for simulating base drift mixed noise signal after integrated filter；
Fig. 5 is that true base floats effect picture of the mixed noise signal after integrated filter；
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, detailed
Carefully describe.Described embodiment is only a part of the embodiments of the present invention.
The technical solution that the present invention solves above-mentioned technical problem is:
The invention proposes a kind of based on wavelet transformation and improves the electrocardiosignal integrated filter method of EEMD.Its core is thought
Think be: for noise behavior in electrocardio, Main Noise Sources are divided into low-frequency disturbance and High-frequency Interference.High-frequency Interference include power frequency and
Myoelectricity interference, design wavelet thresholding method realize the elimination of High-frequency Interference；Low-frequency disturbance, that is, baseline drift interference, for EEMD algorithm
Deficiency, provide the foundation that the size of aid in noise is added in important parameter in EEMD algorithm and set average time determines, then
Baseline drift interference is eliminated using EEMD algorithm is improved.The case where finally existing simultaneously for a variety of interference of electrocardio, devises base
In the comprehensive denoising algorithm of the electrocardiosignal of wavelet transformation and improvement EEMD.The present invention can be improved signal-to-noise ratio, reduce mean square deviation, protect
Hold ecg wave form feature.
Embodiment of the present invention is described in detail with reference to the accompanying drawing.
A kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD, the specific steps of which are as follows:
Step 1: establishing noisy electrocardiosignal model: power frequency, myoelectricity, base being added in 103 relatively pure electrocardiosignals
The interference such as line drift, obtain noisy electrocardiosignal x'；
In general, in electrocardio there are three types of main noise sources: power frequency, myoelectricity and baseline drift interference.Wherein baseline floats
The influence to electrocardiosignal is moved to be particularly acute.The characteristics of for noise, it is as follows to establish electrocardiosignal noise model:
(1) pure electrocardiosignal is chosen (in " MIT-BIH Noise Stress Test Database " database
The sample rate of 103 signals, MLII lead, signal is 360Hz, takes 4096 data points), shown in signal waveform such as attached drawing 3 (a)；
(2) simulate Hz noise: power frequency interference signals using amplitude be 0.12mV, frequency be 50Hz sinusoidal signal, 103
Signal averaging industrial frequency noise postwave is shaped like shown in attached drawing 3 (b)；
(3) simulate myoelectricity interference: myoelectricity interference uses white noise, and 103 Signal averaging myoelectricity noise postwaves are shaped like attached drawing 3
(c) shown in；
(4) baseline drift is interfered: there are two types of form (drift of simulation base and the drifts of true base) for the baseline drift interference of introducing.Simulation
Base drift uses amplitude for 0.2mV, and frequency is the sinusoidal signal of 0.5Hz, and 103 Signal averagings simulate base drift postwave shaped like attached drawing 3 (d)
It is shown.True base, which winnows, takes the base in " MIT-BIH Noise Stress Test Database " database to float interference signal
(bw-noise1), the true base drift postwave of 103 Signal averagings is shaped like shown in attached drawing 3 (e)；
(5) under normal circumstances, collected electrocardiosignal can be simultaneously comprising above several interference, and attached drawing 3 (f) is 103 letters
Waveform after number superposed simulation baseline drift, power frequency and myoelectricity interference；Attached drawing 3 (g) be the true baseline drift of 103 Signal averagings,
Waveform after power frequency and myoelectricity interference.
Step 2: wavelet threshold eliminates high-frequency noise: being decomposed using wavelet basis coif3 to signals and associated noises, use threshold
Value method eliminates high-frequency noise, using soft-threshold mode, selection of threshold function " Rigrsure ", the signal sequence after eliminating high-frequency noise
Column are denoted as x；
Wavelet transformation is as a kind of Time-Frequency Analysis Method, and in the denoising of signal, compression etc. is widely used.
Especially in terms of signal denoising, wavelet transformation plays highly important role.Noise Elimination from Wavelet Transform mainly has following step
It is rapid:
(1) signal decomposition.Firstly the need of using suitable small echo to decompose signal；
(2) threshold value quantizing.It chooses suitable threshold function table and threshold value quantizing is carried out to the component after wavelet decomposition；
(3) signal reconstruction.The reconstruct of signal is carried out to the component by threshold value quantizing.
Wherein second step threshold value quantizing is the key that Wavelet Denoising Method.Threshold value quantizing mainly includes three kinds of situations: (1) defaulting threshold
Value；(2) given threshold value；(3) denoising (wavelet coefficient zero) is forced.
This project is not in further investigated wavelet transformation basic function and the On The Choice of threshold function table, with reference to related article, choosing
More appropriate basic function and threshold function table is taken to carry out the elimination of high-frequency noise.Attached drawing 4 (c) and attached drawing 5 (c) are to remove electrocardio
Waveform diagram after high-frequency noise.
Step 3: obtaining the important parameter of EEMD algorithm: being decomposed using EMD to noisy electrocardiosignal, by first
IMF component obtains the value for improving parameter k and N in EEMD algorithm as high fdrequency component；
It include two highly important parameters in EEMD algorithm: the size and ensemble average of the auxiliary white noise of addition
Number.Under normal conditions, the two parameters are all rule of thumb to be configured, then using EEMD when handling signal,
The selection of the two parameters will necessarily have a huge impact the decomposition result of signal, this does not expect to see, and should
It allows algorithm itself to determine the two parameters, optimizes decomposition result.I.e. improved EEMD algorithm.
The present invention gives with reference to pertinent literature is added any discontinuous signal auxiliary white noise in EEMD method
Sound can be according to criterion.The white noise of addition must meet following condition:
In formula, the radio-frequency component standard deviation of A --- signal and the proportionality coefficient of original signal standard deviation
Under normal circumstances, it takesThe modal overlap phenomenon occurred when signal decomposition can effectively be avoided.
Ensemble average number is another important parameter in EEMD method, it can determine to assist after signal decomposition white
The elimination situation of noise.Z.H.Wu et al., which has studied ensemble average times N and proportionality coefficient k in this method, should meet such as ShiShimonoseki
System:
First IMF component that selection signal obtains after the decomposition of EMD method is as signal radio-frequency component.To obtain
Coefficient A；Then coefficient k is obtained according to formula (1)；Presetting relative error e simultaneously, (it is 1% that e, which under normal conditions, is arranged,
Meet the requirements), according to the available ensemble average times N of formula (2).It is thus available important to two in EEMD algorithm
Parameter.
Step 4: being decomposed using EEMD algorithm is improved: after determining parameter, using improved EEMD algorithm to elimination
Signal sequence x after high-frequency noise is decomposed；
Above-mentioned steps confirmed two important parameters k and N in EEMD algorithm, then be set accordingly in the algorithm
Set, then signal sequence x to be processed decomposed using improved EEMD algorithm, obtained after decomposition several modal components and
Residual volume.
Step 5: obtaining the baseline drift signal extracted: after signal decomposition, average to IMF component and residual volume, root
According to discriminate | mean (IMFi) | > ε chooses component part of the IMF component for the condition that meets as extraction base drift, obtained base
Drift signal is denoted as f；
Step 6: reconstruct obtains the electrocardiosignal after de-noising: subtracting the base drift letter of extraction using the signal x after elimination high frequency
Number f, the signal y=x-f after de-noising can be obtained.Attached drawing 4 (d) is the waveform diagram removed after EGC analog baseline drift；Attached drawing 5
It (d) is the waveform diagram removed after the true baseline drift of electrocardio.Since then, which completes disappearing for electrocardio noise jamming
It removes.
The above embodiment is interpreted as being merely to illustrate the present invention rather than limit the scope of the invention.?
After the content for having read record of the invention, technical staff can be made various changes or modifications the present invention, these equivalent changes
Change and modification equally falls into the scope of the claims in the present invention.
Claims (6)
1. a kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD, which is characterized in that including following step
It is rapid:
Step 1: establishing noisy electrocardiosignal model: 103 electrocardiosignal of MIT-BIH database is chosen, wherein 103 electrocardiosignals
It is influenced by baseline drift than more serious, and the interference including power frequency, myoelectricity, baseline drift is added, obtain noisy electrocardiosignal
x'；
Step 2: wavelet threshold eliminates high-frequency noise: being decomposed, used to noisy electrocardiosignal x' using wavelet basis coif3
Threshold method eliminates high-frequency noise, and the threshold method uses soft-threshold mode, and selection of threshold function " Rigrsure " eliminates high frequency and makes an uproar
Signal sequence after sound is denoted as x；
Step 3: use experience mode decomposition (EMD) decomposes the signal x after eliminating high-frequency noise, it is intrinsic by first
Mode function component obtains the value for improving parameter k and N in EEMD algorithm as high fdrequency component；Wherein k is white noise and signal
The ratio of amplitude standard deviation, N are ensemble average number；
Step 4: being decomposed using EEMD algorithm is improved: after determining parameter, using improved EEMD algorithm to elimination high frequency
Signal sequence x after noise is decomposed；Improved EEMD provides important parameter in EEMD algorithm and the big of aid in noise is added
The foundation that small and set average time determines, the parameter k and N of original EEMD is entirely by being manually set, none foundation determined,
According to discriminate | mean (IMFi) | > ε chooses the IMF component for the condition that meets as the component part for extracting base drift, obtains
Base drift signal is denoted as；
Step 5: obtaining the baseline drift signal extracted: after signal decomposition, average to all IMF components and residual volume, root
According to discriminate | mean (IMFi) | > ε chooses component part of the IMF component for the condition that meets as extraction base drift, obtained base
Drift signal is denoted as f；
Step 6: reconstruct obtains the electrocardiosignal after de-noising: signal f is floated using the base that the signal x after elimination high frequency subtracts extraction,
Signal y=x-f after de-noising can be obtained.
2. a kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD according to claim 1,
It is characterized in that, the step 1 chooses pure electrocardiosignal " MIT-BIH Noise Stress Test Database " data
The sample rate of 103 signals in library, MLII lead, signal is 360Hz, takes 4096 data points.
3. a kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD according to claim 1,
It being characterized in that, the interference including power frequency, myoelectricity, baseline drift is added in the step 1, it specifically includes:
Simulate Hz noise: power frequency interference signals are 0.12mV using amplitude, and frequency is the sinusoidal signal of 50Hz；It is dry to simulate myoelectricity
Disturb: myoelectricity interference uses white noise；
Baseline drift interference: the baseline drift interference of introducing has the drift of simulation base and true base to float two kinds of forms, and simulation base drift uses
Amplitude is 0.2mV, and frequency is the sinusoidal signal of 0.5Hz, and true base, which winnows, takes " MIT-BIH Noise Stress Test
Base in Database " database floats interference signal.
4. a kind of electrocardiosignal integrated filter side based on wavelet transformation and improvement EEMD described in one of -3 according to claim 1
Method, which is characterized in that in the step 2, mainly include three kinds of situations in the threshold value quantizing step of wavelet transformation: (1) defaulting
Threshold value；(2) given threshold value；(3) forcing denoising is wavelet coefficient zero.
5. a kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD according to claim 4,
It is characterized in that, in the step 3, obtains the value for improving parameter k and N in EEMD algorithm, specifically include: the white noise of addition
Following condition must be met:
In formula, the radio-frequency component standard deviation of A --- signal and the proportionality coefficient of original signal standard deviation；It takes
Ensemble average times N and proportionality coefficient k should meet following relationship:
First IMF component that selection signal obtains after the decomposition of EMD method is as signal radio-frequency component, to obtain coefficient
A；Then coefficient k is obtained according to formula (1)；Relative error e is preset according to the available ensemble average times N of formula (2) simultaneously.
6. a kind of electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD according to claim 5,
It is characterized in that, the step 4 is decomposed using EEMD algorithm is improved: after determining parameter, using improved EEMD algorithm pair
Signal sequence x after eliminating high-frequency noise is decomposed；Specific step is as follows:
(1) auxiliary white noise signal+k σ is added in the signal x of removal high-frequency noisexN (t), wherein k be white noise with
The ratio of signal amplitude standard deviation, σxPoor for signal standards, n (t) is normalization white noise, constitutes noise mixture:
X1(t)=x (t)+k σx·n(t) (3)
(2) EMD decomposition is carried out to noise mixture:
Wherein cjIndicate j-th of intrinsic mode function, r that noise mixture generates after EMD is decomposedmIndicate noise mixture warp
Trend term after EMD decomposition, m indicate the quantity of intrinsic mode function after EMD decomposition；
(3) step (1) and (2) is repeated, different white noises is added every time:
Xi(t)=x (t)+kni(t) (5)
Resolve into IMF:
ci,jThe noise mixture for indicating that i-th is added after assisting white noise in the signal x of removal high-frequency noise is decomposed through EMD
J-th of intrinsic mode function, r afterwardsi,mIndicate that the letter after auxiliary white noise is added in i-th in the signal x of removal high-frequency noise
It makes an uproar trend term of the mixture after EMD is decomposed；
(4) n times are repeated, are then averaged to each IMF:
(5) last decomposition result are as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910246351.3A CN110051325A (en) | 2019-03-29 | 2019-03-29 | Electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910246351.3A CN110051325A (en) | 2019-03-29 | 2019-03-29 | Electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110051325A true CN110051325A (en) | 2019-07-26 |
Family
ID=67317811
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910246351.3A Pending CN110051325A (en) | 2019-03-29 | 2019-03-29 | Electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110051325A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111956210A (en) * | 2020-08-27 | 2020-11-20 | 重庆邮电大学 | Electrocardiosignal R wave identification method based on EEMD and signal structure analysis |
| CN112583378A (en) * | 2020-12-04 | 2021-03-30 | 山东省科学院自动化研究所 | Photoelectric sensing signal reconstruction method and system containing baseline drift and high-frequency noise |
| CN112595782A (en) * | 2020-11-17 | 2021-04-02 | 江西理工大学 | Ultrasonic transverse wave trip point identification method and system based on EEMD algorithm |
| CN113180685A (en) * | 2020-12-11 | 2021-07-30 | 上海交通大学 | Electrocardio abnormity discrimination system and method based on morphological filtering and wavelet threshold |
| CN113303809A (en) * | 2021-05-27 | 2021-08-27 | 河北省科学院应用数学研究所 | Method, device, equipment and storage medium for removing baseline drift and high-frequency noise |
| CN114264724A (en) * | 2021-12-23 | 2022-04-01 | 大连海洋大学 | Ultrasonic signal processing method based on CEEMDAN combined wavelet packet threshold |
| CN116304777A (en) * | 2023-04-12 | 2023-06-23 | 中国科学院大学 | Self-adaptive electrocardiosignal denoising method and system based on reference signal during rest |
| CN116509415A (en) * | 2023-04-21 | 2023-08-01 | 山东省人工智能研究院 | Noise reduction method based on unitized morphology features of signal components |
| CN116548928A (en) * | 2023-07-11 | 2023-08-08 | 西安浩阳志德医疗科技有限公司 | Nursing service system based on internet |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104367316A (en) * | 2014-11-13 | 2015-02-25 | 重庆邮电大学 | Electrocardiosignal denoising method based on morphological filtering and lifting wavelet transformation |
| CN108158573A (en) * | 2017-12-26 | 2018-06-15 | 智慧康源（厦门）科技有限公司 | Electrocardiosignal noise-reduction method based on adaptive threshold wavelet transformation |
| CN108294745A (en) * | 2018-03-07 | 2018-07-20 | 武汉大学 | P waves, T wave start-stop point detecting methods and system in multi-lead ECG signal |
| CN108618773A (en) * | 2017-03-15 | 2018-10-09 | 深圳市理邦精密仪器股份有限公司 | A kind of Denoising of ECG Signal, device and a kind of ecg signal acquiring equipment |
| CN108814579A (en) * | 2018-04-16 | 2018-11-16 | 西安交通大学 | A method of electrocardio, breathing combined calculation heart rate variability based on EMD decomposition |
-
2019
- 2019-03-29 CN CN201910246351.3A patent/CN110051325A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104367316A (en) * | 2014-11-13 | 2015-02-25 | 重庆邮电大学 | Electrocardiosignal denoising method based on morphological filtering and lifting wavelet transformation |
| CN108618773A (en) * | 2017-03-15 | 2018-10-09 | 深圳市理邦精密仪器股份有限公司 | A kind of Denoising of ECG Signal, device and a kind of ecg signal acquiring equipment |
| CN108158573A (en) * | 2017-12-26 | 2018-06-15 | 智慧康源（厦门）科技有限公司 | Electrocardiosignal noise-reduction method based on adaptive threshold wavelet transformation |
| CN108294745A (en) * | 2018-03-07 | 2018-07-20 | 武汉大学 | P waves, T wave start-stop point detecting methods and system in multi-lead ECG signal |
| CN108814579A (en) * | 2018-04-16 | 2018-11-16 | 西安交通大学 | A method of electrocardio, breathing combined calculation heart rate variability based on EMD decomposition |
Non-Patent Citations (2)
| Title |
|---|
| 宋美: "基于集合经验模态分解和小波收缩算法的自适应心电信号去噪问题研究", 《生物数学学报》 * |
| 林金朝等: "基于改进EEMD的心电信号基线漂移消除方法", 《数据采集与处理》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111956210A (en) * | 2020-08-27 | 2020-11-20 | 重庆邮电大学 | Electrocardiosignal R wave identification method based on EEMD and signal structure analysis |
| CN111956210B (en) * | 2020-08-27 | 2022-06-03 | 重庆邮电大学 | Electrocardiosignal R wave identification method based on EEMD and signal structure analysis |
| CN112595782A (en) * | 2020-11-17 | 2021-04-02 | 江西理工大学 | Ultrasonic transverse wave trip point identification method and system based on EEMD algorithm |
| CN112595782B (en) * | 2020-11-17 | 2022-07-22 | 江西理工大学 | Ultrasonic transverse wave take-off point identification method and system based on EEMD algorithm |
| CN112583378A (en) * | 2020-12-04 | 2021-03-30 | 山东省科学院自动化研究所 | Photoelectric sensing signal reconstruction method and system containing baseline drift and high-frequency noise |
| CN112583378B (en) * | 2020-12-04 | 2023-07-25 | 山东省科学院自动化研究所 | Photoelectric sensing signal reconstruction method and system containing baseline drift and high-frequency noise |
| CN113180685A (en) * | 2020-12-11 | 2021-07-30 | 上海交通大学 | Electrocardio abnormity discrimination system and method based on morphological filtering and wavelet threshold |
| CN113180685B (en) * | 2020-12-11 | 2022-10-18 | 上海交通大学 | Electrocardio abnormity discrimination system and method based on morphological filtering and wavelet threshold |
| CN113303809B (en) * | 2021-05-27 | 2022-08-26 | 河北省科学院应用数学研究所 | Method, device, equipment and storage medium for removing baseline drift and high-frequency noise |
| CN113303809A (en) * | 2021-05-27 | 2021-08-27 | 河北省科学院应用数学研究所 | Method, device, equipment and storage medium for removing baseline drift and high-frequency noise |
| CN114264724A (en) * | 2021-12-23 | 2022-04-01 | 大连海洋大学 | Ultrasonic signal processing method based on CEEMDAN combined wavelet packet threshold |
| CN116304777A (en) * | 2023-04-12 | 2023-06-23 | 中国科学院大学 | Self-adaptive electrocardiosignal denoising method and system based on reference signal during rest |
| CN116304777B (en) * | 2023-04-12 | 2023-11-03 | 中国科学院大学 | Self-adaptive electrocardiosignal denoising method and system based on reference signal during rest |
| CN116509415A (en) * | 2023-04-21 | 2023-08-01 | 山东省人工智能研究院 | Noise reduction method based on unitized morphology features of signal components |
| CN116509415B (en) * | 2023-04-21 | 2024-01-26 | 山东省人工智能研究院 | Noise reduction method based on unitized morphology features of signal components |
| CN116548928A (en) * | 2023-07-11 | 2023-08-08 | 西安浩阳志德医疗科技有限公司 | Nursing service system based on internet |
| CN116548928B (en) * | 2023-07-11 | 2023-09-08 | 西安浩阳志德医疗科技有限公司 | Nursing service system based on internet |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110051325A (en) | Electrocardiosignal integrated filter method based on wavelet transformation and improvement EEMD | |
| Xiong et al. | ECG signal enhancement based on improved denoising auto-encoder | |
| CN104367316B (en) | Denoising of ECG Signal based on morphologic filtering Yu lifting wavelet transform | |
| Islam et al. | Study and analysis of ecg signal using matlab &labview as effective tools | |
| CN103156599B (en) | Detection method of electrocardiosignal R characteristic waves | |
| CN103405227B (en) | Double-layer morphological filter based electrocardiosignal preprocessing method | |
| CN102697493A (en) | Method for rapidly and automatically identifying and removing ocular artifacts in electroencephalogram signal | |
| CN108338784A (en) | The Denoising of ECG Signal of wavelet entropy threshold based on EEMD | |
| CN110338786A (en) | A kind of identification of epileptiform discharges and classification method, system, device and medium | |
| CN107693004A (en) | Fetal ECG extraction and fetal heart frequency recognition methods based on hilbert conversion | |
| Abdelmounim et al. | Electrocardiogram signal denoising using discrete wavelet transform | |
| Kumar et al. | Performance analysis of different filters for power line interface reduction in ECG signal | |
| Mishra et al. | Denoising of heart sound signal using wavelet transform | |
| Jeyarani et al. | Analysis of noise reduction techniques on QRS ECG waveform-by applying different filters | |
| Zhao et al. | Denoising of ECG signals based on CEEMDAN | |
| Malleswari et al. | An improved denoising of electrocardiogram signals based on wavelet thresholding | |
| Agrawal et al. | Wavelet subband dependent thresholding for denoising of phonocardiographic signals | |
| Zhao et al. | [Retracted] An Early Warning of Atrial Fibrillation Based on Short‐Time ECG Signals | |
| Linsen et al. | Research on ECG signal denoising method based on sparse component analysis | |
| abdullah et al. | Separation enhancement of power line noise from human ECG signal based on stone technique | |
| Haibing et al. | Discrete wavelet soft threshold denoise processing for ECG signal | |
| Jiménez et al. | Performance of a method to generate fetal cardiotachograms using fetal phonocardiography | |
| Tripathi et al. | Statistical analysis of HRV parameters for the detection of Arrhythmia | |
| Strachan et al. | Automated QT analysis that learns from cardiologist annotations | |
| Bing et al. | Hybrid Fourier-Wavelet Heart Magnetic Field Signal Denoising Based on Sparse Representation |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190726 |