CN105701456A - Angular accelerometer signal adaptive denoising method based on wavelet analysis - Google Patents
Angular accelerometer signal adaptive denoising method based on wavelet analysis Download PDFInfo
- Publication number
- CN105701456A CN105701456A CN201610005949.XA CN201610005949A CN105701456A CN 105701456 A CN105701456 A CN 105701456A CN 201610005949 A CN201610005949 A CN 201610005949A CN 105701456 A CN105701456 A CN 105701456A
- Authority
- CN
- China
- Prior art keywords
- signal
- noise
- wavelet
- threshold value
- detail coefficients
- 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
-
- 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)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Complex Calculations (AREA)
Abstract
The present invention provides an angular accelerometer signal adaptive denoising method based on wavelet analysis. Compared with a conventional wavelet analysis denoising method, the method of the present invention comprises the steps of calculating the difference quotient of the wavelet energy entropies and the energy entropies of a signal with noise and a noisy signal to determine an optimal wavelet decomposition level of the signal with noise, applying the median filtering to evaluate a noise width and a noise standard deviation, and combining a 3 sigma criterion to set a threshold value to thereby carry out the adaptive filtering. At the aspect of the threshold value selection, the method has better effects by being compared with a conventional heuristic sure criterion, a maximum and minimum value criterion and a penalty criterion. The above methods have better effects on the denoising processing of a periodic signal and a high-dynamic signal of a molecule type liquid-ring angular accelerometer.
Description
Technical field
The present invention relates to a kind of angular accelerometer signal adaptive denoising method based on wavelet analysis, belong to signal processing technology field。
Background technology
Angular movement is widely present in sea, land and sky, sky every field, is one of most basic motion of nature。In angular movement dynamic characteristic sign, angular acceleration can characterize the high-order characteristic of angular movement more direct, more rapid, more accurately。Along with new and high technology is in the extensive use of military field, what future war faced is combined operation comprehensive, high intensity, multi-class weapon。Therefore under the complex working conditions such as many disturbances, varying load, big overload, in high precision, height measures angular acceleration signal dynamically, highly reliably, the general operation effectiveness of armament systems can be improved, improve heavy-duty machinery, the task performance of the industrial equipments such as precision instrument, improves the Operating ettectiveness of the vehicle such as aircraft, underwater sailing body。
Molecule-type liquid-ring angular accelerometer is as the sensor of a kind of direct measurement angular acceleration signal, interfacial electric double layer effect is utilized to carry out the axial angular acceleration signal of sensitizing input, and export the signal of telecommunication being directly proportional to this angular acceleration signal, compared to tradition rate gyroscope data are carried out electronics or the obtained angular acceleration signal differential calculation of electromagnetism differential, solve angular acceleration error to amplify and to postpone the problems such as delayed, reliability and capacity of resisting disturbance higher。But molecule-type liquid-ring angular accelerometer still can be subject to the restriction of the complex environment factors such as own material structure, carrier mechanical vibration and friction, output signal even can produce distortion with much noise。Especially angular accelerometer is usually operated under the complex environment of many disturbances, varying load, output signal more contains high-frequency oscillation signal useful in a large number sometimes, it is necessary to be filtered just obtaining signal more accurate, reliable to carry out the data analysis work in later stage。
At present, the Study of filtering algorithm for inertial sensor output signals such as angular accelerometers is concentrated mainly on Kalman filtering algorithm。Although Kalman filtering algorithm is the algorithm of a kind of optimal estimation, but system is first set up model by its needs, utilizes linear system state equation to carry out state observation and estimation。For molecule-type liquid-ring angular accelerometer, domestic not yet obtain clear and definite system model, therefore cannot carry out signal processing by Kalman filtering algorithm。And adopt average, the median filtering algorithm commonly used, and also cannot take into account the smooth of noise and the reservation to local high-frequency signal even if regulating sliding window simultaneously, self-adaptive solution degree is low, causes that signal reliability declines。Visible, for the process of molecule-type liquid-ring angular accelerometer use in complex environment and high dynamic signal but without proposing a kind of effective solution。
Summary of the invention
In view of this, the present invention provides a kind of angular accelerometer signal adaptive denoising method based on wavelet analysis, based on Wavelet Analysis Theory, high dynamic characteristic according to angular accelerometer signal carries out adaptive-filtering, can smooth noise effectively, the high-frequency oscillation signal that can remain with again, makes signal to noise ratio be improved。
A kind of angular accelerometer signal adaptive denoising method based on wavelet analysis, comprising:
Step one: the angular accelerometer output signal obtained is carried out length judgement, and this angular accelerometer output signal is grandfather tape noise cancellation signal and is designated as XN, if length is more than 1000, then takes front 1000 as the sample XN of signals with noisesam；If length is less than 1000, then with the grandfather tape noise cancellation signal XN sample XN being signals with noisesam；Then the sample XN to signals with noisesamIt is normalized computing and obtains signals with noise X, simultaneously according to Mmax=log2(length(XNsam)) calculate maximum decomposition level number Mmax；And generate the white noise sequence sample N that length is 1000sam, and obtain noise sequence N after it is normalized computing；
Selecting wavelet basis function is that db3 wavelet basis carries out m layer wavelet decomposition and initial m=1；
Step 2: utilize db3 wavelet basis function respectively signals with noise X and noise sequence N to be carried out m layer wavelet decomposition, obtains approximation coefficient cAm and detail coefficients cDm, and calculates the M shell wavelet energy entropy WEE of signals with noise X and noise sequence N respectivelyXAnd WEEN；
Step 3: according toCalculate the difference coefficient T of layering wavelet energy entropy, if m=MmaxOr T > 10%, then stopping calculating, the Decomposition order now obtained deducts one for best Decomposition order, if best Decomposition order is M；If T < 10%, then updating m value is m+1, returns step 2 and continues to calculate；
Step 4: grandfather tape noise cancellation signal XN is carried out medium filtering and obtains signal Xmid, and with XN and XmidSubtract each other and obtain Noise Estimation sequenceAskAbsolute value after average again acquisitionHalf width W, simultaneously calculate Noise Estimation sequenceStandard deviationFinally according toWithObtain threshold value upper bound gateHWith lower bound gateL；
Step 5: grandfather tape noise cancellation signal XN is carried out M shell discrete wavelet transformation with db3 wavelet basis function, obtain approximation coefficient cAM and detail coefficients cD1, cD2 ... cDM, detail coefficients is compared with setting threshold value, retains in detail coefficients more than threshold value upper bound gateHAnd less than threshold value lower bound gateLDetail coefficients in element value, it is thus achieved that the detail coefficients cD1 ', cD2 ' after process ... cDM '；
Step 6: with each layer detail coefficients cD1 ', cD2 ' after approximation coefficient cAM and threshold process ... cDM ' carries out one-dimensional wavelet inverse transformation, obtains final denoised signal
Beneficial effect:
The present invention is based on Wavelet Analysis Theory, and it is carried out signal processing by the high dynamic characteristics of signals of binding molecule type liquid-ring angular accelerometer。This algorithm is different from the Kalman filtering algorithm dependence to sensing system model, it is different from the median filtering algorithm dependence to window size, and the restriction that Mean Filtering Algorithm is to high dynamic signal process aspect, signal characteristic from actual application environment, smooth noise effectively, the high-frequency oscillation signal that can remain with again, makes signal to noise ratio be improved。
The present invention is compared to traditional wavelet analysis denoising method, propose the best wavelet decomposition number of plies determining signals with noise by calculating signals with noise and the wavelet energy entropy of noise signal and the difference coefficient of Energy-Entropy, use medium filtering that noise width and noise criteria difference are estimated, threshold value is set, thus carrying out adaptive-filtering in conjunction with 3 σ criterions。In threshold value is chosen, compared to traditional heuristic sure criterion, Min-max criterion and penalty criterion, effect is even better。Above method all achieves good effect for the cyclical signal of molecule-type liquid-ring angular accelerometer and the noise reduction process of high dynamic signal。
Accompanying drawing explanation
Fig. 1 is Whole Work Flow figure of the present invention；
Fig. 2 is the flow chart that in the present invention, wavelet decomposition number of plies self adaptation is determined；
Fig. 3 is the flow chart that in the present invention, small echo threshold value is determined；
Fig. 4 is each original emulation signal graph being not added with white noise in the present invention；
Fig. 5 obtains each signals with noise figure that signal to noise ratio is 15dB after Additive White Noise in the present invention；
Fig. 6 obtains each signals with noise figure that signal to noise ratio is 10dB after Additive White Noise in the present invention；
Fig. 7 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Block signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 15dB；
Fig. 8 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Block signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 10dB；
Fig. 9 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Bumps signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 15dB；
Figure 10 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Bumps signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 10dB；
Figure 11 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Heavysines signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 15dB；
Figure 12 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Heavysines signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 10dB；
Figure 13 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Doppler signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 15dB；
Figure 14 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe Doppler signal de-noising design sketch of Min-max threshold method, original signal to noise ratio is 10dB；
Figure 15 is molecule-type liquid-ring angular accelerometer output voltage frequency be 5HZ, signal intensity is the original signal sequence figure of 940.5mV；
Figure 16 is for adopting Threshold and heuristic threshold selection method, generic threshold value in the present inventionThe angular accelerometer signal de-noising design sketch of Min-max threshold method；
Detailed description of the invention
For the technological means making the present invention realize, creation characteristic, reach purpose and effect and be apparent to, below in conjunction with detailed description of the invention, the present invention is expanded on further。
With reference to Fig. 1, this instantiation is by the following technical solutions: a kind of based on wavelet analysis, for the self-adaptive solution method of molecule-type liquid-ring angular accelerometer signal, the steps include: 1, angular acceleration meter signal carries out small echo layering Energy-Entropy and calculates, obtain wavelet decomposition the best number of plies M, signal is carried out M shell discrete wavelet transformation, obtains the wavelet coefficient on each decomposition scale；2, by coupling calculating wavelet decomposition threshold value with median filtering algorithm, by threshold value, wavelet coefficient is carried out threshold value quantizing process；3, last, the wavelet coefficient that utilization processed carries out one-dimensional wavelet inverse transformation, completes denoising。
1, the self adaptation of the wavelet decomposition number of plies is determined
In traditional engineer applied, the determination of the discrete wavelet transformation number of plies is typically with following methods: by signals with noise carries out m=1 layer scattering wavelet decomposition, the detail coefficients obtained after decomposing is carried out albefaction inspection, if the detail coefficients sequence of this layer is white noise sequence, then Decomposition order m+1, until then the discrete wavelet transformation number of plies is M=m-1 layer when detail coefficients can not be checked by albefaction。But when angular accelerometer is applied in the environment of many disturbances, its actual signal also can present and have the high behavioral characteristics of dependency with white noise, now cannot simply by the carrying out albefaction inspection and distinguish actual signal and noise signal of detail coefficients after decomposing。
Therefore, the present invention is according to the E.T.Jaynes principle of maximum entropy proposed in nineteen fifty-seven, proposing a kind of wavelet energy entropy algorithm to determine to the self adaptation realizing the wavelet decomposition number of plies, this kind of method judges Decomposition order based on the difference of noise signal Yu actual signal Energy distribution。When angular accelerometer is applied in the environment of many disturbances, detail coefficients after its output signal decomposition can contain the high dynamic part of actual signal, their Energy distribution of detail coefficients after contrast white noise signal decomposition has obvious difference, can judge Decomposition order relatively reliable, exactly by setting the threshold value of this difference。
Wavelet energy entropy algorithm is from the energy of signal, and relevant to the complexity of signal and Decomposition order, specific algorithm is as follows:
Definition Y (t) represents the angular accelerometer signal sequence that length is n, then its energy is:
Every layer of detail coefficients after then Y (t) signal being carried out M shell discrete wavelet transformation is Yi(t), i=1,2 ..., M, the proportionality coefficient P always accounting for detail coefficients energy summation is accounted for for i-th layer of wavelet energy entropy WEE (i) and every layer of detail coefficients energyjIt is given by:
Concrete operating procedure is given below:
1) as shown in Figure 2, it is necessary first to obtain angular accelerometer signal。What the present invention adopted is the molecule-type liquid-ring angular accelerometer of N0.33 RESEARCH INSTITUTE OF THE THIRD ACADEMY OF CASIC's development, it is installed on the turntable of adjustable speed, given turntable is with the sine voltage signal of different frequency, thus getting many groups voltage signal that under different frequency, angular accelerometer exports。
The required variable of definition is as follows:
XN={XNi| i=1,2 ..., n} represents the grandfather tape noise cancellation signal sequence that length is n；
XNsam={ XNi| i=1,2 ..., m, m=min{n, 1000}} represents sample strip noise cancellation signal, wherein, and min{n, 1000} represents, if grandfather tape noise cancellation signal length n is more than 1000, then takes front 1000 data of grandfather tape noise cancellation signal as sample sequence, otherwise take grandfather tape noise cancellation signal total data as sample sequence；
X={Xi| i=1,2 ..., m} represents sample strip noise cancellation signal is normalized the signal sequence of gained after computing；
Nsam={ Ni| i=1,2 ..., m} represents the white noise sequence that generation length is identical with sample strip noise cancellation signal；
N={Ni| i=1,2 ..., m} represents white noise sequence is normalized the signal sequence of gained after computing；
S={Si| i=1,2 ..., n} represents authentic signal sequence；
Represent the signal sequence after denoising；
MmaxRepresent the maximum decomposable asymmetric choice net number of plies；
M={Mi| i=1,2 ..., MmaxRepresent the actual layer number that signal is carried out discrete wavelet transformation；
For discrete wavelet transformation method, length is the signal of n, and its maximum Decomposition order is given by:
Mmax=log2(length(XNsam))(4)
In order to improve calculating speed, reduce amount of calculation, it is stipulated that maximum decomposition level number is less than 9 layers, then can be calculated signals with noise length length (XN) < 1024 by (4) formula。For the angular accelerometer signal XN obtained, if its length is more than 1000, then take front 1000 data as the sample data calculating best Decomposition order；If its length is less than 1000, then takes total data and be calculated。Thus obtain the sample of signal X for determining Decomposition ordersam, and its maximum decomposition level number MmaxIt it is 9 layers。
2) generate and sample of signal XsamThe white noise sequence sample N that length is identicalsam, and with following formula to XsamAnd NsamIt is normalized simultaneously:
3) signal X and N is carried out M=1 layer wavelet decomposition, formula (2), (3) the wavelet energy entropy WEE of X can be calculatedXWavelet energy entropy WEE with NN；
4) calculating the difference coefficient of the layering Energy-Entropy of X and the N of M shell discrete wavelet transformation gained, formula is as follows:
5) we judge whether to obtain best Decomposition order M by the setting threshold value T notable difference detecting signal X and noise N layering Energy-Entropy。If T < 0.1, then Decomposition order M=M+1, skip to step 3) proceed discrete wavelet transformation；If T >=0.1, now stop computing, obtain final best Decomposition order M=M-1。
2, the determination of wavelet decomposition threshold value
In order to make the actual signal of angular accelerometer better be separated with noise signal, it is achieved preferably denoising effect, except obtaining the rational wavelet decomposition number of plies, choosing of wavelet decomposition threshold value is also most important。
When isolating noise from the detail coefficients of wavelet decomposition, it is necessary to the intensity of noise is reasonably estimated, thus ensureing to the full extent the actual signal part in detail coefficients to be retained。Medium filtering is a kind of common nonlinear properties smoothing technique, and it is theoretical based on sequencing statistical, impulsive noise is had and good filters effect, the profile of signal can be protected and get rid of big burr while effectively suppressing noise。By using medium filtering, it is possible to quickly and accurately estimate the profile of actual signal to a certain extent, then subtract each other with grandfather tape noise cancellation signal, it may be achieved the estimation to noise intensity。
Therefore, invention defines this concept of average half width W of noise, its value is the meansigma methods of noise intensity absolute value, and computing formula is as follows:
Wherein,Represent the estimation noise intensity sequence that length is n。
Further, we can calculate this statistical property of standard deviation sigma of noise, sets wavelet decomposition threshold value thresholding in conjunction with 3 σ rules。When angular accelerometer is in the working environment of many disturbances heavy load, its output signal detail coefficients after carrying out discrete wavelet transformation can be mixed with part high dynamically and the actual signal of high intensity, in order to retain actual signal to greatest extent, filtering noise, can after calculating noise half width, this datum line is applied 3 σ rules as threshold value thresholding, the detail coefficients of the "abnormal" intensity outside beyond white noise strength range is retained, remaining detail coefficients substantially belongs to noise signal within the scope of probability statistics, it is possible to it is carried out zero-setting operation。
Concrete operation step that wavelet threshold determine is given below:
1) as shown in Figure 3, the required variable of definition is as follows:
XN={XNi| i=1,2 ..., n} represents the signals with noise sequence that length is n；
XM={XMi| i=1,2 ..., n} represents that signals with noise passes through the signal sequence obtained after medium filtering；
Represent and estimate noise intensity sequence；
WNRepresent the average half width estimating noise intensity sequence；
gateLRepresent the upper bound of threshold value, gateHRepresent the lower bound of threshold value；
First, selection window is sized to 5% length of sequence length n, and signals with noise XN is carried out medium filtering, obtains filtered sequence X M；
2) calculating estimation noise intensity sequence, computing formula is as follows:
3) calculateHalf width WNAnd standard deviationComputing formula is as follows:
4) finally obtaining lower bound and the upper bound of threshold value, computing formula is as follows:
So far, the determination that angular acceleration meter signal carries out the threshold value thresholding of discrete wavelet transformation is completed。
3, threshold process and one-dimensional wavelet reconstruction
The present invention self adaptation can determine the threshold value thresholding of the discrete wavelet transformation number of plies and detail coefficients by above Double Step algorithm, complete the denoising of angular acceleration meter signal, also needing by threshold process and the operation of one-dimensional wavelet inverse transformation two step, concrete calculation procedure is as follows:
1) first, with db3 wavelet basis function, grandfather tape noise cancellation signal XN is carried out M shell discrete wavelet transformation, obtains approximation coefficient cAM and detail coefficients cD1, cD2 ... cDM；
2) according to the threshold value upper bound calculated in 2 and lower bound, each layer detail coefficients is carried out threshold process, between (gateL,gateH) between detail coefficients all carry out zero-setting operation, the part beyond threshold value is retained, thus the detail coefficients cD1 ', cD2 ' after being processed ... cDM '；
3) with each layer the detail coefficients cD1 ', cD2 ' after approximation coefficient cAM and threshold process ... cDM ' carries out one-dimensional wavelet inverse transformation, obtains final denoised signal。
4, instance analysis
1) nonstationary random signal emulation experiment
The present invention adopts Matlab establishment based on the angular acceleration signal denoising method program of wavelet analysis, first choose Blocks, Bumps, Heavysine, Doopler in Matlab to emulate the nonstationary random signal similar to angular accelerometer signal characteristic and test, as shown in Figure 4。The sampled point of four kinds of signals is 2048, and on signal, superposition white Gaussian noise obtains signal to noise ratio respectively is the signals and associated noises of 10dB and 15dB, as shown in Figure 5, Figure 6。Apply and proposed by the invention determine that the Decomposition order of each nonstationary random signal is such as shown in table 1 table 2 based on wavelet energy entropy algorithm self adaptation:
Table 1 signal to noise ratio is the optimal self-adaptive Decomposition order of each noisy nonstationary random signal when being 10dB
Table 2 signal to noise ratio is the optimal self-adaptive Decomposition order of each noisy nonstationary random signal when being 15db
Daubechies wavelet function from MATLAB wavelet library and in Symlets wavelet function, chooses db3 small echo and processes。
Gauge signal noise reduction generally utilizes signal to noise ratio snr, mean square error MSE。Signal XN (i) after S without noise cancellation signal (i) superposition white Gaussian noise signal N (i), XN (i)=S (i)+N (i), i=1,2,3......n, wherein n is sampled point number。For the signal after denoising, then signal-to-noise ratio computation formula is formula (15):
Then the signal-to-noise ratio computation formula after noise reduction is as follows:
Mean square error is defined as:
Primary signal is more high with the root-mean-square error signal to noise ratio more little, signal of denoised signal, then denoised signal is just closer to real primary signal, and denoising effect is also just corresponding better。Use conventional threshold values method: heuristic threshold selection method, generic threshold valueFour kinds of random stochastic signals of non-stationary are carried out noise reduction process by Min-max threshold method and thresholding method proposed by the invention respectively, result is such as shown in table 3, table 4, it is known that the thresholding method that the present invention proposes can obtain higher signal to noise ratio and less mean square error。
Table 3 emulates signal de-noising result (original signal to noise ratio is 10dB)
Table 4 emulates signal de-noising result (original signal to noise ratio is 15dB)
4 kinds of thresholding method are used respectively such as shown in Fig. 6 Figure 14, the result of 4 kinds of signals can intuitively to be found out, four class signals are carried out noise reduction and are substantially better than other several conventional threshold values functions by the wavelet decomposition Threshold that present invention determine that。
2) molecule-type liquid-ring angular accelerometer actual signal denoising effect
Present invention employs the molecule-type liquid-ring angular accelerometer that N0.33 RESEARCH INSTITUTE OF THE THIRD ACADEMY OF CASIC develops, it is installed on the turntable of adjustable speed, given turntable is with the sine voltage signal of different frequency, thus getting many groups voltage signal that under different frequency, angular accelerometer exports。It is carried out denoising by the signal sequence choosing one of which output voltage frequency to be 5HZ, signal intensity be 940.5mV, its true primary signal is as shown in figure 15, use the method that the present invention proposes, calculating the best wavelet decomposition number of plies is 4 layers, and after carrying out contrast denoising with conventional threshold values system of selection, effect is as shown in figure 16。
In sum, these are only presently preferred embodiments of the present invention, be not intended to limit protection scope of the present invention。All within the spirit and principles in the present invention, any amendment of making, equivalent replacement, improvement etc., should be included within protection scope of the present invention。
Claims (1)
1. the angular accelerometer signal adaptive denoising method based on wavelet analysis, it is characterised in that including:
Step one: the angular accelerometer output signal obtained is carried out length judgement, and this angular accelerometer output signal is grandfather tape noise cancellation signal and is designated as XN, if length is more than 1000, then takes front 1000 as the sample XN of signals with noisesam；If length is less than 1000, then with the grandfather tape noise cancellation signal XN sample XN being signals with noisesam；Then the sample XN to signals with noisesamIt is normalized computing and obtains signals with noise X, simultaneously basis: Mmax=log2(length(XNsam)) calculate maximum decomposition level number Mmax；And generate the white noise sequence sample N that length is 1000sam, and obtain noise sequence N after it is normalized computing；
Selecting wavelet basis function is that db3 wavelet basis carries out m layer wavelet decomposition and initial m=1；
Step 2: utilize db3 wavelet basis function respectively signals with noise X and noise sequence N to be carried out m layer wavelet decomposition, obtains approximation coefficient cAm and detail coefficients cDm, and calculates the M shell wavelet energy entropy WEE of signals with noise X and noise sequence N respectivelyXAnd WEEN；
Step 3: according toCalculate the difference coefficient T of layering wavelet energy entropy, if m=MmaxOr T > 10%, then stopping calculating, the Decomposition order now obtained deducts one for best Decomposition order, if best Decomposition order is M；If T < 10%, then updating m value is m+1, returns step 2 and continues to calculate；
Step 4: grandfather tape noise cancellation signal XN is carried out medium filtering and obtains signal Xmid, and with XN and XmidSubtract each other and obtain Noise Estimation sequenceAskAbsolute value after average again acquisitionHalf width W, simultaneously calculate Noise Estimation sequenceStandard deviationFinally according toWithObtain threshold value upper bound gateHWith lower bound gateL；
Step 5: grandfather tape noise cancellation signal XN is carried out M shell discrete wavelet transformation with db3 wavelet basis function, obtain approximation coefficient cAM and detail coefficients cD1, cD2 ... cDM, detail coefficients is compared with setting threshold value, retains in detail coefficients more than threshold value upper bound gateHAnd less than threshold value lower bound gateLDetail coefficients in element value, it is thus achieved that the detail coefficients cD1 ', cD2 ' after process ... cDM '；
Step 6: with each layer detail coefficients cD1 ', cD2 ' after approximation coefficient cAM and threshold process ... cDM ' carries out one-dimensional wavelet inverse transformation, obtains final denoised signal
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610005949.XA CN105701456B (en) | 2016-01-05 | 2016-01-05 | A kind of angular accelerometer signal adaptive denoising method based on wavelet analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610005949.XA CN105701456B (en) | 2016-01-05 | 2016-01-05 | A kind of angular accelerometer signal adaptive denoising method based on wavelet analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105701456A true CN105701456A (en) | 2016-06-22 |
| CN105701456B CN105701456B (en) | 2019-10-25 |
Family
ID=56226201
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610005949.XA Active CN105701456B (en) | 2016-01-05 | 2016-01-05 | A kind of angular accelerometer signal adaptive denoising method based on wavelet analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105701456B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106441288A (en) * | 2016-08-31 | 2017-02-22 | 北斗时空信息技术（北京）有限公司 | Adaptive wavelet denoising method for accelerometer |
| CN108898117A (en) * | 2018-06-30 | 2018-11-27 | 防灾科技学院 | A kind of self-adapting random abnormal signal extracting method for sliding threshold value |
| CN109190220A (en) * | 2018-08-22 | 2019-01-11 | 宁波洁程汽车科技有限公司 | A kind of engine air-tightness diagnostic method and system based on wavelet analysis |
| CN109738519A (en) * | 2019-01-04 | 2019-05-10 | 国网四川省电力公司广安供电公司 | A kind of denoising method of transformer high-voltage bushing lead ultrasound detection |
| CN110113279A (en) * | 2019-05-05 | 2019-08-09 | 哈尔滨工程大学 | A kind of mobile frequency hopping underwater sound communication Doppler factor estimation method |
| CN110287853A (en) * | 2019-06-20 | 2019-09-27 | 清华大学 | A kind of Transient Signal Denoising based on wavelet decomposition |
| CN114048771A (en) * | 2021-11-09 | 2022-02-15 | 西安电子科技大学 | Time sequence data abnormal value processing method based on adaptive threshold stationary wavelet transformation |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006130010A2 (en) * | 2005-06-03 | 2006-12-07 | Epsis As | Method for processing sampled data |
| CN101425176A (en) * | 2008-12-09 | 2009-05-06 | 中国科学院长春光学精密机械与物理研究所 | Image wavelet de-noising method based on median filter |
| CN104251865A (en) * | 2013-06-26 | 2014-12-31 | 中南大学 | Method for detecting visible foreign matters in medical medicaments based on affinity propagation clustering |
| CN104346516A (en) * | 2013-08-09 | 2015-02-11 | 中国科学院沈阳自动化研究所 | Wavelet denoising optimal decomposition level selection method of laser-induced breakdown spectroscopy |
-
2016
- 2016-01-05 CN CN201610005949.XA patent/CN105701456B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006130010A2 (en) * | 2005-06-03 | 2006-12-07 | Epsis As | Method for processing sampled data |
| CN101425176A (en) * | 2008-12-09 | 2009-05-06 | 中国科学院长春光学精密机械与物理研究所 | Image wavelet de-noising method based on median filter |
| CN104251865A (en) * | 2013-06-26 | 2014-12-31 | 中南大学 | Method for detecting visible foreign matters in medical medicaments based on affinity propagation clustering |
| CN104346516A (en) * | 2013-08-09 | 2015-02-11 | 中国科学院沈阳自动化研究所 | Wavelet denoising optimal decomposition level selection method of laser-induced breakdown spectroscopy |
Non-Patent Citations (3)
| Title |
|---|
| YAN-FANG SANG ET AL: "Entropy-Based Method of Choosing the Decomposition Level in Wavelet Threshold De-noising", 《ENTROPY》 * |
| YING MA ET AL: "A novel method based on adaptive median filtering and wavelet transform in noise images", 《2011 IEEE 3RD INTERNATIONAL CONFERENCE ON COMMUNICATION SOFTWARE AND NETWORKS》 * |
| 王雯 等: "飞行数据改进中值滤波和自适应小波降噪", 《计算机仿真》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106441288A (en) * | 2016-08-31 | 2017-02-22 | 北斗时空信息技术（北京）有限公司 | Adaptive wavelet denoising method for accelerometer |
| CN106441288B (en) * | 2016-08-31 | 2019-12-20 | 北斗时空信息技术（北京）有限公司 | Self-adaptive wavelet denoising method for accelerometer |
| CN108898117A (en) * | 2018-06-30 | 2018-11-27 | 防灾科技学院 | A kind of self-adapting random abnormal signal extracting method for sliding threshold value |
| CN109190220A (en) * | 2018-08-22 | 2019-01-11 | 宁波洁程汽车科技有限公司 | A kind of engine air-tightness diagnostic method and system based on wavelet analysis |
| CN109190220B (en) * | 2018-08-22 | 2023-04-07 | 宁波洁程汽车科技有限公司 | Engine air tightness diagnosis method and system based on wavelet analysis |
| CN109738519A (en) * | 2019-01-04 | 2019-05-10 | 国网四川省电力公司广安供电公司 | A kind of denoising method of transformer high-voltage bushing lead ultrasound detection |
| CN109738519B (en) * | 2019-01-04 | 2021-08-17 | 国网四川省电力公司广安供电公司 | Denoising method for ultrasonic detection of lead of high-voltage bushing of transformer |
| CN110113279A (en) * | 2019-05-05 | 2019-08-09 | 哈尔滨工程大学 | A kind of mobile frequency hopping underwater sound communication Doppler factor estimation method |
| CN110113279B (en) * | 2019-05-05 | 2021-09-28 | 哈尔滨工程大学 | Mobile frequency hopping underwater acoustic communication Doppler factor estimation method |
| CN110287853A (en) * | 2019-06-20 | 2019-09-27 | 清华大学 | A kind of Transient Signal Denoising based on wavelet decomposition |
| CN114048771A (en) * | 2021-11-09 | 2022-02-15 | 西安电子科技大学 | Time sequence data abnormal value processing method based on adaptive threshold stationary wavelet transformation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105701456B (en) | 2019-10-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105701456A (en) | Angular accelerometer signal adaptive denoising method based on wavelet analysis | |
| CN103163050B (en) | Metal abrasive particle detection method in lubrication oil system based on electromagnetic induction signal | |
| CN110388570B (en) | VMD-based self-adaptive noise reduction method and application thereof in water supply pipeline leakage positioning | |
| CN109359567B (en) | Parameterized transmission path analysis method based on improved wavelet threshold denoising | |
| CN106441288A (en) | Adaptive wavelet denoising method for accelerometer | |
| CN104807534A (en) | Equipment natural vibration mode self-learning recognition method based on online vibration data | |
| CN112686096A (en) | Rolling bearing fault diagnosis method based on multi-scale diffusion entropy and VPMCD | |
| CN112163536B (en) | Wavelet threshold function denoising method based on particle swarm optimization improvement | |
| Li et al. | An EEMD‐Based Denoising Method for Seismic Signal of High Arch Dam Combining Wavelet with Singular Spectrum Analysis | |
| Liu | Additive white Gaussian noise level estimation based on block SVD | |
| CN112882115B (en) | Magnetotelluric signal denoising method and system based on GWO optimized wavelet threshold | |
| CN113326616A (en) | Slow variable coarse error measurement resistant fault-tolerant Kalman filtering method | |
| Shi et al. | The VMD-scale space based hoyergram and its application in rolling bearing fault diagnosis | |
| CN111982489B (en) | Weak fault feature extraction method for selectively integrating improved local feature decomposition | |
| CN111427096A (en) | Data quality evaluation and filtering processing method for full tensor gravity gradiometer | |
| Sun et al. | Application of a novel improved adaptive CYCBD method in gearbox compound fault diagnosis | |
| CN114813123A (en) | Rolling bearing weak fault diagnosis method based on PSO-VMD-MCKD | |
| CN108398260A (en) | The fast evaluation method of gear-box instantaneous angular velocity based on mixing probabilistic method | |
| CN113657244B (en) | Fan gearbox fault diagnosis method and system based on improved EEMD and speech spectrum analysis | |
| CN104283529A (en) | High order volume Kalman filtering method for square root with unknown measurement noise variance | |
| Wang et al. | Estimating dynamic motion parameters with an improved wavelet thresholding and inter-scale correlation | |
| CN114282441A (en) | Friction kinematic pair equivalent model and modeling method thereof | |
| CN117828447A (en) | Aeroengine bearing fault diagnosis method based on spectrogram wavelet transformation and multi-scale permutation entropy | |
| CN110657807B (en) | Indoor positioning displacement measurement method for detecting discontinuity based on wavelet transformation | |
| Liu et al. | A Morphology Filter‐Assisted Extreme‐Point Symmetric Mode Decomposition (MF‐ESMD) Denoising Method for Bridge Dynamic Deflection Based on Ground‐Based Microwave Interferometry |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |