CN108491355A - A kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet - Google Patents
A kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet Download PDFInfo
- Publication number
- CN108491355A CN108491355A CN201810110081.9A CN201810110081A CN108491355A CN 108491355 A CN108491355 A CN 108491355A CN 201810110081 A CN201810110081 A CN 201810110081A CN 108491355 A CN108491355 A CN 108491355A
- Authority
- CN
- China
- Prior art keywords
- noise
- signal
- ceemd
- imf
- wavelet packet
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
- G06F17/148—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
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention discloses a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet, this method carries out mode decomposition first with CEEMD algorithms to signal, obtain a series of intrinsic mode functions and a trend term, secondly soft threshold de-noising is carried out to noise dominant mode in intrinsic mode function, the adaptation rule noise reduction of unbiased possibility predication principle is carried out to signal dominant mode, the capacity of decomposition for finally utilizing wavelet packet analysis fine is to the further denoising of signal.CEEMD algorithms are that two pairs of opposite white noises are added in original signal, EMD decomposition is carried out respectively and result is averaged, it can effectively overcome the problems, such as end effect and modal overlap, and wavelet and wavelet packet decomposes all further noise reductions under conditions of not adding aid in noise.The carried noise-reduction method of the present invention is more preferably than traditional denoising method performance.It the composite can be widely applied to fault in material signal processing.
Description
Technical field
The present invention relates to flaw echoes processing technology field, especially a kind of ultrasound based on CEEMD and wavelet packet
Signal de-noising method.
Background technology
In Ultrasonic NDT, noise is often mixed in useful signal, thus the key for obtaining defect information is
How noise reduction is carried out to echo-signal.Non-linear, non-stationary due to echo-signal, traditional Fourier transformation is difficult to accurately
It identifies useful signal, carries out effective denoising.Improvement of the wavelet transformation as Fourier transformation has preferable Time Frequence Analysis
Ability, the disadvantages of this method are that denoising effect is affected by the selection of wavelet basis function, Decomposition order and threshold value, and choose
Principle is often determined by experience.Wavelet package transforms are the new developments of wavelet transformation, can be provided for signal a kind of more fine
Analysis method, the high frequency section that wavelet transformation does not segment further is decomposed, to improve the processing capacity to signal,
But there are still the disadvantages identical with wavelet transformation.Empirical mode decomposition (EMD) is a kind of time frequency analysis new method, can be preferable
Non-linear and non-stationary signal is handled, and adaptivity is stronger, is a series of consolidating for arrangements from high frequency to low frequency by signal decomposition
The shortcomings that having mode function (IMF) and a residual error, overcoming wavelet threshold denoising method, however it also has its shortcoming, such as
Modal overlap, end effect etc..Gather empirical mode decomposition (EEMD), using the zero mean characteristic of white noise, to original signal
In repeatedly add different white noises, can offset white noise using multiple averaging, obtain a series of IMF, it is so effective that overcome
EMD decompose present in modal overlap phenomenon.Complementary set empirical mode decomposition (CEEMD) method based on EEMD is by two
Opposite white noise is added in original signal, carry out EMD decomposition respectively and by result averagely obtain final
IMF, this method further alleviate modal overlap problem, while keeping decomposition result more thorough.
The studies above analyzes the feature from small echo, WAVELET PACKET DECOMPOSITION to EMD, EEMD and CEEMD algorithm respectively, however
All there are respective advantage and disadvantage in these methods, and how to find an optimal algorithm or algorithm fusion mode, can be to super
Acoustical signal flaw echo can retain useful signal to a greater extent while effectively removing noise signal, still need to further
Research.
Invention content
It is provided the technical problem to be solved by the present invention is to overcome the deficiencies in the prior art a kind of based on CEEMD and small echo
The ultrasonic signal noise-reduction method of packet while the present invention effectively removes noise signal, can retain more useful signals.
The present invention uses following technical scheme to solve above-mentioned technical problem：
According to a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet proposed by the present invention, include the following steps：
Step 1: establishing noisy ultrasonic signal mathematical model；
Step 2: generating signals and associated noises using the established mathematical model of step 1, and carries out CEEMD and decompose noise reduction, it will
Signals and associated noises are divided into n intrinsic mode function IMF, acquire the auto-correlation function of each IMF；
Step 3: according to the characteristic of auto-correlation function, critical mode is judged, it is two classes that IMF, which is divided to, to noise dominant mould
State carries out soft threshold de-noising, and the adaptation rule noise reduction of unbiased possibility predication principle is carried out to signal dominant mode；
Step 4: all IMF are reconstructed, the signal after reconstruct is chosen into wavelet basis function and Decomposition order progress is small
Wave packet noise reduction, obtains the useful signal after noise reduction.
As a kind of ultrasonic signal noise-reduction method side of advanced optimizing based on CEEMD and wavelet packet of the present invention
Case carries out the detailed process that CEEMD decomposes noise reduction in step 2 to signals and associated noises：
1) the auxiliary white noise of a pair of isometric, established standards difference positive and negative white noise composition, is added into signals and associated noises,
Generate two new signals；
2) EMD decomposition, is carried out respectively to two new signals in step 1), obtains two groups of IMF components, every group of n is a
IMF；
3) times N, is decomposed according to the CEEMD of setting, repeats n times step 1) and step 2), what is be added every time is all one group
Random auxiliary white noise；
4) obtained 2N group IMF components are averaged to get to the n IMF generated after CEEMD is decomposed.
As a kind of ultrasonic signal noise-reduction method side of advanced optimizing based on CEEMD and wavelet packet of the present invention
Case, in step 1), the amplitude of the positive and negative white noise being added to signals and associated noises is that the standard deviation of signals and associated noises is multiplied by coefficient p, 0<p<
1。
As a kind of ultrasonic signal noise-reduction method side of advanced optimizing based on CEEMD and wavelet packet of the present invention
Case, when the CEEMD of setting decomposes times N=100, the value range of p is 0.01<p<0.5.
As a kind of ultrasonic signal noise-reduction method side of advanced optimizing based on CEEMD and wavelet packet of the present invention
Case, in step 2, to the normalized autocorrelation functions ρ of each IMF components obtained after CEEMD is decomposedxCalculating it is as follows：
I) auto-correlation function reflects degree of correlation of the same signal sequence between the value of different moments, definition
For：
Rx(t1,t2)=E [x (t1)x(t2)]
Wherein, x (t1)、x(t2) be signal x (t) in t1、t2The value at moment, E [x (t1)x(t2)] indicate mathematic expectaion.
Ii) normalized autocorrelation functions ρ (x) expression formulas are：
ρ (x)=Rx(t1,t2)/Rx(0)
Rx(0) indicate that obtained IMF is in the auto-correlation function value of synchronization after the step is decomposed.
As a kind of ultrasonic signal noise-reduction method side of advanced optimizing based on CEEMD and wavelet packet of the present invention
IMF is divided in step 3 as follows for the processing procedure of two classes by case：
A) according to the characteristic of normalized autocorrelation functions, judge critical mode；
B) after selecting a wavelet basis function and determining Decomposition order, soft threshold de-noising is carried out to noise dominant mode, together
When followed by selection one wavelet basis function and determine Decomposition order after, to signal dominant mode carry out unbiased possibility predication principle
Adaptation rule noise reduction.
As a kind of ultrasonic signal noise-reduction method side of advanced optimizing based on CEEMD and wavelet packet of the present invention
Case, wavelet basis function are selected from the small wave systems of Daubechies.
The present invention has the following technical effects using above technical scheme is compared with the prior art：
(1) present invention fast and effeciently can carry out noise reduction to defect ultrasonic signal；
(2) while the present invention effectively removes noise signal, more useful signals can be retained；
(3) the carried noise-reduction method of the present invention is more preferably than traditional denoising method performance.Material is the composite can be widely applied to lack
Fall into signal processing.
Description of the drawings
Fig. 1 is the algorithm flow chart of the present invention.
Fig. 2 is the algorithm flow chart of EMD.
Fig. 3 is the frequency domain distribution schematic diagram of three layers of WAVELET PACKET DECOMPOSITION.
Fig. 4 is the normalized autocorrelation functions figure of random noise and general noise；Wherein, (a) is random noise signal,
(b) be random signal normalized autocorrelation functions, be (c) band make an uproar and rule general signal, (d) be band make an uproar and rule it is general
The normalized autocorrelation functions of signal.
Fig. 5 is Doppler signals, the Doppler signals that signal-to-noise ratio is 7dB and its signal after noise reduction；Wherein, (a) is
Doppler signals (b) are Doppler signals that signal-to-noise ratio is 7dB, are (c) that signal in (b) passes through inventive algorithm noise reduction
Signal afterwards.
Fig. 6 is the IMF component maps obtained after CEEMD is decomposed.
Fig. 7 is the auto-correlation function of each IMF components.
Fig. 8 is multiple ultrasonic signal, the multiple ultrasonic signal that signal-to-noise ratio is 7dB and its signal after noise reduction；Wherein, (a)
It is multiple ultrasonic signal, is (b) the multiple ultrasonic signal that signal-to-noise ratio is 7dB, is (c) that the signal in (b) passes through inventive algorithm
Signal after noise reduction.
Specific implementation mode
Technical scheme of the present invention is described in further detail below in conjunction with the accompanying drawings：
The present invention proposes a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet, as shown in Figure 1：First with
CEEMD algorithms carry out mode decomposition to signal, obtain a series of intrinsic mode functions and a trend term, secondly to intrinsic mode
Noise dominant mode carries out soft threshold de-noising in function, and the adaptive rule of unbiased possibility predication principle are carried out to signal dominant mode
Then noise reduction, the capacity of decomposition for finally utilizing wavelet packet analysis fine is to the further denoising of signal.
CEEMD algorithms are that two pairs of opposite white noises are added in original signal in method proposed by the present invention, respectively
It carries out EMD decomposition simultaneously result is averaged, can effectively overcome the problems, such as end effect and modal overlap, and wavelet and wavelet packet
Decompose all further noise reductions under conditions of not adding aid in noise.The carried noise-reduction method of the present invention is than traditional denoising method performance
More preferably.It the composite can be widely applied to fault in material signal processing.
Technical scheme of the present invention is further elaborated below by specific embodiment：
The specific implementation step of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet in the present embodiment is as follows：
Step 1：Establish noisy ultrasonic signal mathematical model, including in common MATLAB Doppler signals and
Multiple ultrasonic signal, as shown in (a) in (a) and Fig. 8 in Fig. 5.
When to plastic products defects detection, the ultrasonic wave sent out by ultrasound emission probe, by material surface, fault location
And bottom surface can reflect to form ultrasonic echo, common ultrasonic echo can be divided into single echo and multiple echo.Single echo ultrasound is believed
Number mathematical model S (t) it is as follows：
Wherein β is amplitude, and e is math constant, is the truth of a matter of natural logrithm function, and α is bandwidth factor, and t is the time, and τ is
Time of arrival (toa), fcIt is the centre frequency of signal,It is phase.
It is as follows that multiple ultrasonic echo mathematical model y (t) is can be generalized to according to single echo mathematical model：
Wherein, U is ultrasonic echo tuple, s (θr, t) and indicate r weight ultrasonic echos, θr=[βr,αr,τr,fcr,φr], t tables
Show that time, v (t) are offsets.Choose three groups of different θrRespectively：[3.2,25,3.5,5,1.8], [1,25,5,8,5] with
And [2,25,5,8,5], multiple ultrasonic signal is obtained, as shown in (a) in Fig. 8.
Step 2：It sets Parameters on Defect signal appropriate and carries out CEEMD decomposition noise reductions, signals and associated noises are decomposed into n
IMF(IMF1, IMF2..., IMFn), acquire the auto-correlation function of each IMF.The parameter set decomposes times N and auxiliary as CEEMD
White noise acoustic amplitude p is helped, wherein 0<p<1, as N=100, the value range of p is 0.01<p<0.5.
CEEMD is a kind of algorithm being improved based on EMD and EEMD, including following steps：
1) the auxiliary white noise of a pair of positive and negative white noise composition is added into original signal, generates two new signals：
Wherein：S is original signal, and R is the auxiliary white noise being added, M1、M2The letter after positive and negative white noise is respectively added
Number, therefore 2N groups IMF set can be obtained.
2) EMD decomposition is carried out respectively to two new signals in step 1), obtains two groups of IMF components, every group of n IMF,
clgG-th of IMF component of first of signal is denoted as.
3) times N is decomposed according to the CEEMD of setting, repeats n times step 1) and step 2), be added every time be all one group with
The auxiliary white noise of machine.
4) obtained 2N groups IMF is averaged, obtains n final IMF：
The flow chart of EMD algorithms is as shown in Fig. 2, steps are as follows：
I) maximum point and minimum point for finding out given original signal x (t) are fitted respectively using cubic spline function
The envelope up and down of x (t).
Ii envelope mean value m (t) up and down) is acquired, being subtracted with original signal is worth c (t)=x (t)-m (t).
Iii) in general, by ii) obtained c (t) is unsatisfactory for the condition of IMF, enable x (t)=c (t), repeat i) ii) straight
Meet the condition of IMF to obtained c (t), c (t) at this time is the first rank IMF, i.e. IMF1。
Iv) IMF is subtracted with original signal x (t)1, obtain single order remainder r1(t), r1(t)=x (t)-IMF1。
V) x (t)=r is enabled1(t), i)-iv is repeated), it is referred to as residual until cannot decomposing again until remainder is monotonic function
Difference finally obtains q ranks IMF and a residual error, i.e.,
Usual IMF1Including the most high frequency section of signal, while being also the most part of noise, residual error rqRepresent signal
Trend term.
By the study found that the noise energy in signal is concentrated mainly on the smaller modal components of exponent number, there will necessarily be one
IMF can be divided into two parts by a critical mode by noise dominant and signal are leading.For the Bu Tong spy of useful signal and noise
Property, it can find this critical mode using auto-correlation function.Auto-correlation function statistically reflects same signal sequence and exists
Degree of correlation between the value of different moments is a kind of statistically independent of time domain, thus frequently as signal analysis field
Important evidence.It is defined as：
Rx(t1,t2)=E [x (t1)x(t2)] (8)
Wherein, x (t1)、x(t2) be the signal x (t) in t1、t2The value at moment, E [x (t1)x(t2)] indicate the mathematics phase
It hopes.
Usually reflect this degree using normalized autocorrelation functions in engineer application, so-called normalization refers to will the amount of having
Guiding principle formula is transformed to dimensionless formula, between signal sequence is zoomed to -1 to 1, normalized autocorrelation functions ρxExpression formula be：
ρx=Rx(t1,t2)/Rx(0) (9)
Rx(0) indicate that obtained IMF is in the auto-correlation function value of synchronization after the step is decomposed.
All there is the signal sequence of white noise weak dependence and uncertainty, the sequence of general signal to exist at any time
Rule, therefore the auto-correlation function of white noise is maximum at zero, zero is all decayed to immediately at other points, and general signal exists
Auto-correlation function reaches maximum value at zero, and zero is not decayed to immediately at remaining point, but is influenced by correlation, slow wave
It is dynamic to decline, as shown in figure 4, Fig. 4 is the normalized autocorrelation functions figure of random noise and general noise；Wherein, (a) in Fig. 4
It is random noise signal, (b) in Fig. 4 is random signal normalized autocorrelation functions, and (c) in Fig. 4 is that band is made an uproar and rule
General signal, (d) in Fig. 4 are that band is made an uproar and the normalized autocorrelation functions of the general signal of rule.
By taking signal-to-noise ratio shown in fig. 5 is the Doppler signal decomposition of 7dB as an example, Fig. 5 is that Doppler signals, signal-to-noise ratio are
Signal after the Doppler signals and its noise reduction of 7dB；Wherein, (a) in Fig. 5 is Doppler signals, and (b) in Fig. 5 is letter
It makes an uproar than the Doppler signals for 7dB, (c) in Fig. 5 is signal of the signal after inventive algorithm noise reduction in (b)；It should
Signal decomposition is a series of IMF as shown in fig. 6, the auto-correlation function for then acquiring each IMF components is as shown in Figure 7.
Step 3：According to auto-correlation function characteristic, critical mode IMF is judgedk, soft-threshold is carried out to noise dominant mode
Noise reduction carries out signal dominant mode the adaptation rule noise reduction of unbiased possibility predication principle.
For shown in Fig. 7, critical point k=5 can be obtained according to autocorrelation function graph and auto-correlation function characteristic, selected
The small wave systems of Daubechies are taken as wavelet basis function and determine that suitable Decomposition order carries out soft-threshold to 1-4 rank IMF components
Noise reduction, while choosing the small wave systems of Daubechies as wavelet basis function and determining suitable Decomposition order to 5-8 rank IMF components
Carry out the adaptation rule noise reduction of unbiased possibility predication principle.Other kinds of signal is similarly handled using the program.
Step 4：All IMF are reconstructed, the signal after reconstruct is chosen into suitable wavelet basis function and Decomposition order
Carry out wavelet-packet noise reduction.
Wavelet package transforms are the improvement and development of wavelet transformation, it can provide finer analysis ability for signal, to small
There is no the high frequency section of subdivision to be decomposed again in wave conversion, can further distinguish useful signal and noise signal, three layers
The frequency domain distribution schematic diagram of WAVELET PACKET DECOMPOSITION is as shown in figure 3, wherein F indicates that original signal, H indicate that the signal obtained after decomposing is high
Frequency part, L then indicate low frequency part.
The selection formula of threshold value is in WAVELET PACKET DECOMPOSITION：
Wherein w is the length of handled signal.
Signal reconstruction is carried out to the handling result of step 3, is then handled, is chosen using the scheme of the step
The small wave systems of Daubechies simultaneously determine that suitable Decomposition order carries out wavelet-packet noise reduction, obtain the final handling result of the algorithm such as
Shown in (c) in Fig. 5.
It is Signal to Noise Ratio (SNR) and mean square error RMSE, formula point that the performance measure index after noise reduction is carried out using the algorithm
It is not as follows：
Wherein, SNR is bigger, shows that useful signal ratio is bigger in result, and noise reduction is better, RMSE then with this phase
Instead, table 1 is the performance indicator that the algorithm carries out the Doppler signals that signal-to-noise ratio is 7dB noise reduction.
| Performance indicator | The method of the present invention |
| SNR/dB | 20.217 8 |
| RMSE | 0.211 4 |
1 signal-to-noise ratio of table is the Doppler signal de-noising performance indicators of 7dB
In conclusion the algorithm is equally good to the treatment effect of multiple ultrasonic signal, Fig. 8 is multiple ultrasonic signal, letter
It makes an uproar than the signal after the multiple ultrasonic signal and its noise reduction for 7dB；Wherein, (a) in Fig. 8 is multiple ultrasonic signal, in Fig. 8
(b) it is multiple ultrasonic signal that signal-to-noise ratio is 7dB, (c) in Fig. 8 is signal in (b) after inventive algorithm noise reduction
Signal；As shown in (c) in Fig. 8, different noises is added, a series of places are carried out to noisy signal using the method for the invention
Reason, performance indicator SNR and RMSE is as shown in table 2 and table 3.
More echo-signal SNR indexs of different signal-to-noise ratio are added in table 2
| White noise/dB is added | The method of the present invention |
| -1 | 1.917 4 |
| 3 | 5.5335 |
| 7 | 7.833 6 |
| 10 | 8.439 3 |
More echo-signal RMSE indexs of different signal-to-noise ratio are added in table 3
| White noise/dB is added | The method of the present invention |
| -1 | 0.194 8 |
| 3 | 0.143 3 |
| 7 | 0.102 7 |
| 10 | 0.095 5 |
The above, the only specific implementation mode in the present invention, but scope of protection of the present invention is not limited thereto, appoints
What is familiar with the people of the technology within the technical scope disclosed by the invention, it will be appreciated that expects transforms or replaces, and should all cover
Within the scope of the present invention, therefore, the scope of protection of the invention shall be subject to the scope of protection specified in the patent claim.
Claims (7)
1. a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet, which is characterized in that include the following steps：
Step 1: establishing noisy ultrasonic signal mathematical model；
Step 2: generating signals and associated noises using the established mathematical model of step 1, and carries out CEEMD and decompose noise reduction, it will be noisy
Signal is divided into n intrinsic mode function IMF, acquires the auto-correlation function of each IMF；
Step 3: according to the characteristic of auto-correlation function, judge critical mode, by IMF be divided to be two classes, to noise dominant mode into
Row soft threshold de-noising carries out signal dominant mode the adaptation rule noise reduction of unbiased possibility predication principle；
Step 4: all IMF are reconstructed, the signal after reconstruct is chosen into wavelet basis function and Decomposition order carries out wavelet packet
Noise reduction obtains the useful signal after noise reduction.
2. a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet according to claim 1, which is characterized in that
The detailed process that CEEMD decomposes noise reduction is carried out to signals and associated noises in step 2：
1) the auxiliary white noise of a pair of isometric, established standards difference positive and negative white noise composition, is added into signals and associated noises, generates
Two new signals；
2) EMD decomposition, is carried out respectively to two new signals in step 1), obtains two groups of IMF components, every group of n IMF；
3) times N, is decomposed according to the CEEMD of setting, repeats n times step 1) and step 2), what is be added every time is all one group random
Auxiliary white noise；
4) obtained 2N group IMF components are averaged to get to the n IMF generated after CEEMD is decomposed.
3. a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet according to claim 2, which is characterized in that
In step 1), the amplitude of the positive and negative white noise being added to signals and associated noises is that the standard deviation of signals and associated noises is multiplied by coefficient p, 0<p<1.
4. a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet according to claim 3, which is characterized in that
When the CEEMD of setting decomposes times N=100, the value range of p is 0.01<p<0.5.
5. a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet according to claim 1, which is characterized in that
In step 2, to the normalized autocorrelation functions ρ of each IMF components obtained after CEEMD is decomposedxCalculating it is as follows：
I) auto-correlation function reflects degree of correlation of the same signal sequence between the value of different moments, is defined as：
Rx(t1,t2)=E [x (t1)x(t2)]
Wherein, x (t1)、x(t2) be signal x (t) in t1、t2The value at moment, E [x (t1)x(t2)] indicate mathematic expectaion.
Ii) normalized autocorrelation functions ρ (x) expression formulas are：
ρ (x)=Rx(t1,t2)/Rx(0)
Rx(0) indicate that obtained IMF is in the auto-correlation function value of synchronization after the step is decomposed.
6. a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet according to claim 5, which is characterized in that
It is divided to the processing procedure for two classes as follows IMF in step 3：
A) according to the characteristic of normalized autocorrelation functions, judge critical mode；
B) after selecting a wavelet basis function and determining Decomposition order, soft threshold de-noising is carried out to noise dominant mode, while again
Then after selecting a wavelet basis function and determination Decomposition order, oneself of unbiased possibility predication principle is carried out to signal dominant mode
Adapt to regular noise reduction.
7. a kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet according to claim 6, which is characterized in that
Wavelet basis function is selected from the small wave systems of Daubechies.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810110081.9A CN108491355A (en) | 2018-02-05 | 2018-02-05 | A kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810110081.9A CN108491355A (en) | 2018-02-05 | 2018-02-05 | A kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108491355A true CN108491355A (en) | 2018-09-04 |
Family
ID=63344578
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810110081.9A Pending CN108491355A (en) | 2018-02-05 | 2018-02-05 | A kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108491355A (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109543599A (en) * | 2018-11-20 | 2019-03-29 | 贵州电网有限责任公司 | A kind of method of transformer fault traveling wave noise reduction |
| CN109567743A (en) * | 2018-10-24 | 2019-04-05 | 加康康健有限公司 | A kind of signal reconfiguring method based on EMD, device, terminal device and storage medium |
| CN109630908A (en) * | 2019-01-23 | 2019-04-16 | 常州大学 | A kind of pipeline leakage positioning method of multiple noise reduction |
| CN109798920A (en) * | 2019-02-18 | 2019-05-24 | 兰州交通大学 | Based on the MEMS gyro stochastic error modeling filtering method for improving EMD |
| CN110096998A (en) * | 2019-04-29 | 2019-08-06 | 天津大学 | A kind of decomposition method extracted for machining surface shape characteristic |
| CN110322364A (en) * | 2019-06-19 | 2019-10-11 | 山东大学 | A kind of short-term photovoltaic power generation prediction technique and system |
| CN110399846A (en) * | 2019-07-03 | 2019-11-01 | 北京航空航天大学 | A kind of gesture identification method based on multichannel electromyography signal correlation |
| CN110487916A (en) * | 2019-08-24 | 2019-11-22 | 东北电力大学 | A kind of noise-reduction method suitable for the dirt ultrasound detection echo-signal that exchanges heat |
| CN110658257A (en) * | 2019-10-10 | 2020-01-07 | 天津科技大学 | Method for identifying defects of aircraft cable insulating layer based on wavelet packet analysis |
| CN110688964A (en) * | 2019-09-30 | 2020-01-14 | 哈尔滨工程大学 | Wavelet threshold and EMD combined denoising method based on sparse decomposition |
| CN111161171A (en) * | 2019-12-18 | 2020-05-15 | 三明学院 | Blasting vibration signal baseline zero drift correction and noise elimination method, device, equipment and system |
| CN111553308A (en) * | 2020-05-11 | 2020-08-18 | 成都亿科康德电气有限公司 | Reconstruction method of partial discharge signal of power transformer |
| CN111795931A (en) * | 2020-07-06 | 2020-10-20 | 武汉理工大学 | Reconstruction extraction method for laser ultrasonic defect detection diffraction echo signal |
| CN112528857A (en) * | 2020-12-10 | 2021-03-19 | 上海海事大学 | EEMD-based noise reduction method for reciprocating friction vibration signal |
| CN112697887A (en) * | 2020-12-08 | 2021-04-23 | 江苏科技大学 | Ultrasonic detection defect qualitative identification method based on neural network |
| CN112964793A (en) * | 2021-02-09 | 2021-06-15 | 常州大学 | Intelligent pipeline pitting damage identification method based on image identification and acoustic identification |
| CN113075706A (en) * | 2021-03-25 | 2021-07-06 | 上海海洋大学 | GNSS-R based snow depth inversion method and application thereof |
| CN113238190A (en) * | 2021-04-12 | 2021-08-10 | 大连海事大学 | Ground penetrating radar echo signal denoising method based on EMD combined wavelet threshold |
| CN113534006A (en) * | 2021-07-11 | 2021-10-22 | 太原理工大学 | Single-phase earth fault line selection method based on CEEMD and autocorrelation threshold denoising |
| CN113569395A (en) * | 2021-07-16 | 2021-10-29 | 沈阳工业大学 | Signal noise reduction method for low-frequency ultrasonic thoracic cavity imaging |
| CN114219088A (en) * | 2021-11-02 | 2022-03-22 | 中北大学 | Decomposition method of aliasing shock wave signal |
| 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 |
| CN115438301A (en) * | 2022-10-26 | 2022-12-06 | 北京建筑大学 | Equipment operation trend prediction method based on ICEEMDAN quadratic decomposition coupling informar model |
| CN117744893A (en) * | 2024-02-19 | 2024-03-22 | 西安热工研究院有限公司 | Wind speed prediction method and system for energy storage auxiliary black start |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103091096A (en) * | 2013-01-23 | 2013-05-08 | 北京信息科技大学 | Extraction method for early failure sensitive characteristics based on ensemble empirical mode decomposition (EEMD) and wavelet packet transform |
| US8632465B1 (en) * | 2009-11-03 | 2014-01-21 | Vivaquant Llc | Physiological signal denoising |
| CN103902844A (en) * | 2014-04-24 | 2014-07-02 | 国家电网公司 | Transformer vibration signal de-noising method based on EEMD kurtosis threshold value |
| US8858450B2 (en) * | 2006-05-16 | 2014-10-14 | The Research Foundation Of State University Of New York | Photoplethysmography apparatus and method employing high resolution estimation of Time-Frequency Spectra |
| CN104408694A (en) * | 2014-11-27 | 2015-03-11 | 北京理工大学 | Denoising method for nonlocal average image based on soft threshold |
| CN104615877A (en) * | 2015-01-28 | 2015-05-13 | 辽宁工程技术大学 | Method for conducting signal denoising based on wavelet packet |
| CN104655425A (en) * | 2015-03-06 | 2015-05-27 | 重庆大学 | Bearing fault classification diagnosis method based on sparse representation and LDM (large margin distribution machine) |
-
2018
- 2018-02-05 CN CN201810110081.9A patent/CN108491355A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8858450B2 (en) * | 2006-05-16 | 2014-10-14 | The Research Foundation Of State University Of New York | Photoplethysmography apparatus and method employing high resolution estimation of Time-Frequency Spectra |
| US8632465B1 (en) * | 2009-11-03 | 2014-01-21 | Vivaquant Llc | Physiological signal denoising |
| CN103091096A (en) * | 2013-01-23 | 2013-05-08 | 北京信息科技大学 | Extraction method for early failure sensitive characteristics based on ensemble empirical mode decomposition (EEMD) and wavelet packet transform |
| CN103902844A (en) * | 2014-04-24 | 2014-07-02 | 国家电网公司 | Transformer vibration signal de-noising method based on EEMD kurtosis threshold value |
| CN104408694A (en) * | 2014-11-27 | 2015-03-11 | 北京理工大学 | Denoising method for nonlocal average image based on soft threshold |
| CN104615877A (en) * | 2015-01-28 | 2015-05-13 | 辽宁工程技术大学 | Method for conducting signal denoising based on wavelet packet |
| CN104655425A (en) * | 2015-03-06 | 2015-05-27 | 重庆大学 | Bearing fault classification diagnosis method based on sparse representation and LDM (large margin distribution machine) |
Non-Patent Citations (2)
| Title |
|---|
| 喻伟等: "基于CEEMD和小波包变换的重力数据信噪分离方法", 《软件》 * |
| 张樯等: "基于改进的EMD超声信号降噪方法研究", 《南京邮电大学学报( 自然科学版)》 * |
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109567743A (en) * | 2018-10-24 | 2019-04-05 | 加康康健有限公司 | A kind of signal reconfiguring method based on EMD, device, terminal device and storage medium |
| CN109543599A (en) * | 2018-11-20 | 2019-03-29 | 贵州电网有限责任公司 | A kind of method of transformer fault traveling wave noise reduction |
| CN109630908A (en) * | 2019-01-23 | 2019-04-16 | 常州大学 | A kind of pipeline leakage positioning method of multiple noise reduction |
| CN109798920A (en) * | 2019-02-18 | 2019-05-24 | 兰州交通大学 | Based on the MEMS gyro stochastic error modeling filtering method for improving EMD |
| CN110096998A (en) * | 2019-04-29 | 2019-08-06 | 天津大学 | A kind of decomposition method extracted for machining surface shape characteristic |
| CN110096998B (en) * | 2019-04-29 | 2023-05-12 | 天津大学 | Decomposition method for extracting machined surface morphology features |
| CN110322364A (en) * | 2019-06-19 | 2019-10-11 | 山东大学 | A kind of short-term photovoltaic power generation prediction technique and system |
| CN110399846A (en) * | 2019-07-03 | 2019-11-01 | 北京航空航天大学 | A kind of gesture identification method based on multichannel electromyography signal correlation |
| CN110487916A (en) * | 2019-08-24 | 2019-11-22 | 东北电力大学 | A kind of noise-reduction method suitable for the dirt ultrasound detection echo-signal that exchanges heat |
| CN110688964A (en) * | 2019-09-30 | 2020-01-14 | 哈尔滨工程大学 | Wavelet threshold and EMD combined denoising method based on sparse decomposition |
| CN110658257A (en) * | 2019-10-10 | 2020-01-07 | 天津科技大学 | Method for identifying defects of aircraft cable insulating layer based on wavelet packet analysis |
| CN111161171A (en) * | 2019-12-18 | 2020-05-15 | 三明学院 | Blasting vibration signal baseline zero drift correction and noise elimination method, device, equipment and system |
| CN111161171B (en) * | 2019-12-18 | 2023-04-07 | 三明学院 | Blasting vibration signal baseline zero drift correction and noise elimination method, device, equipment and system |
| CN111553308A (en) * | 2020-05-11 | 2020-08-18 | 成都亿科康德电气有限公司 | Reconstruction method of partial discharge signal of power transformer |
| CN111795931A (en) * | 2020-07-06 | 2020-10-20 | 武汉理工大学 | Reconstruction extraction method for laser ultrasonic defect detection diffraction echo signal |
| CN112697887A (en) * | 2020-12-08 | 2021-04-23 | 江苏科技大学 | Ultrasonic detection defect qualitative identification method based on neural network |
| CN112528857A (en) * | 2020-12-10 | 2021-03-19 | 上海海事大学 | EEMD-based noise reduction method for reciprocating friction vibration signal |
| CN112964793A (en) * | 2021-02-09 | 2021-06-15 | 常州大学 | Intelligent pipeline pitting damage identification method based on image identification and acoustic identification |
| CN112964793B (en) * | 2021-02-09 | 2024-01-09 | 常州大学 | Intelligent pipeline pitting damage identification method based on image identification and acoustic identification |
| CN113075706A (en) * | 2021-03-25 | 2021-07-06 | 上海海洋大学 | GNSS-R based snow depth inversion method and application thereof |
| CN113238190A (en) * | 2021-04-12 | 2021-08-10 | 大连海事大学 | Ground penetrating radar echo signal denoising method based on EMD combined wavelet threshold |
| CN113238190B (en) * | 2021-04-12 | 2023-07-21 | 大连海事大学 | Ground penetrating radar echo signal denoising method based on EMD combined wavelet threshold |
| CN113534006B (en) * | 2021-07-11 | 2022-12-27 | 太原理工大学 | Single-phase earth fault line selection method based on CEEMD and autocorrelation threshold denoising |
| CN113534006A (en) * | 2021-07-11 | 2021-10-22 | 太原理工大学 | Single-phase earth fault line selection method based on CEEMD and autocorrelation threshold denoising |
| CN113569395B (en) * | 2021-07-16 | 2023-08-22 | 沈阳工业大学 | Signal noise reduction method for low-frequency ultrasonic thoracic imaging |
| CN113569395A (en) * | 2021-07-16 | 2021-10-29 | 沈阳工业大学 | Signal noise reduction method for low-frequency ultrasonic thoracic cavity imaging |
| CN114219088B (en) * | 2021-11-02 | 2022-09-27 | 中北大学 | Decomposition method of aliasing shock wave signal |
| CN114219088A (en) * | 2021-11-02 | 2022-03-22 | 中北大学 | Decomposition method of aliasing shock wave signal |
| 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 |
| CN115438301A (en) * | 2022-10-26 | 2022-12-06 | 北京建筑大学 | Equipment operation trend prediction method based on ICEEMDAN quadratic decomposition coupling informar model |
| CN115438301B (en) * | 2022-10-26 | 2023-06-20 | 北京建筑大学 | Equipment operation trend prediction method based on ICEEMDAN (information and communication technology) secondary decoupling index model |
| CN117744893A (en) * | 2024-02-19 | 2024-03-22 | 西安热工研究院有限公司 | Wind speed prediction method and system for energy storage auxiliary black start |
| CN117744893B (en) * | 2024-02-19 | 2024-05-17 | 西安热工研究院有限公司 | Wind speed prediction method and system for energy storage auxiliary black start |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108491355A (en) | A kind of ultrasonic signal noise-reduction method based on CEEMD and wavelet packet | |
| Raj et al. | Denoising algorithm using wavelet for underwater signal affected by wind driven ambient noise | |
| Varghese et al. | Direct strain estimation in elastography using spectral cross-correlation | |
| Lazaro et al. | Influence of thresholding procedures in ultrasonic grain noise reduction using wavelets | |
| Dass | Speckle noise reduction of ultrasound images using BFO cascaded with wiener filter and discrete wavelet transform in homomorphic region | |
| CN108954020B (en) | Pipeline positioning method | |
| CN109580787B (en) | Ultrasonic echo denoising method for ultrasonic detection of transformer high-voltage bushing lead | |
| CN112697887B (en) | Ultrasonic detection defect qualitative identification method based on neural network | |
| CN110151175A (en) | Surface electromyogram signal noise-eliminating method based on CEEMD and improvement wavelet threshold | |
| CN109871733A (en) | A kind of adaptive sea clutter signal antinoise method | |
| CN103778921A (en) | Method for eliminating nonuniform noise in speech collected by radar | |
| CN110501158A (en) | Needing machine transmission mechanism vibration signal characteristics extracting method | |
| CN111795931A (en) | Reconstruction extraction method for laser ultrasonic defect detection diffraction echo signal | |
| CN112380934A (en) | Cable partial discharge signal self-adaptive wavelet denoising method based on wavelet entropy and sparsity | |
| CN114264724A (en) | Ultrasonic signal processing method based on CEEMDAN combined wavelet packet threshold | |
| CN115840879A (en) | Additive manufacturing component ultrasonic detection echo signal processing method and system | |
| CN105929029B (en) | One kind is for method for processing noise in SH Guided Wave NDT Technique | |
| CN116698398A (en) | Gear fault feature extraction method based on CEEMDAN sub-threshold noise reduction and energy entropy | |
| Jin et al. | An adaptive ultrasonic backscattered signal processing technique for instantaneous characteristic frequency detection | |
| Mao et al. | Noise suppression and flaw detection of ultrasonic signals via empirical mode decomposition | |
| Zhang et al. | An Efficient Porcine Acoustic Signal Denoising Technique Based on EEMD‐ICA‐WTD | |
| Ou et al. | Frame-based time-scale filters for underwater acoustic noise reduction | |
| CN103971012B (en) | Lamb wave signal denoising method based on John Saris model and fractional differentiation | |
| Lu et al. | The application of an improved wavelet threshold denoising method in heart sound signal | |
| Pardo et al. | Shift invariant wavelet denoising of ultrasonic traces |
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: 20180904 |