CN111650654B - Ground magnetic resonance signal peak noise elimination method combining EMD and WT algorithms - Google Patents
Ground magnetic resonance signal peak noise elimination method combining EMD and WT algorithms Download PDFInfo
- Publication number
- CN111650654B CN111650654B CN202010401183.3A CN202010401183A CN111650654B CN 111650654 B CN111650654 B CN 111650654B CN 202010401183 A CN202010401183 A CN 202010401183A CN 111650654 B CN111650654 B CN 111650654B
- Authority
- CN
- China
- Prior art keywords
- component
- spike noise
- modal
- noise
- magnetic resonance
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
- G01V3/38—Processing data, e.g. for analysis, for interpretation, for correction
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
- G01V3/14—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation operating with electron or nuclear magnetic resonance
Abstract
The invention relates to a ground magnetic resonance signal peak noise eliminating method combining an EMD algorithm and a WT algorithm. According to the method, effective elimination of continuous multiple peak noises in a short time can be realized only by processing the magnetic resonance signals acquired once, so that the inversion accuracy is improved, and the measurement time is saved.
Description
Technical Field
The invention belongs to the field of a ground Nuclear Magnetic Resonance (SNMR) data preprocessing method, and particularly relates to a method for eliminating spike noise of a ground Magnetic Resonance signal based on an EMD (empirical mode decomposition) algorithm and a WT (WT-WT) algorithm.
Background
The magnetic resonance underground water detection technology is the only existing non-invasive geophysical method capable of directly detecting underground water. The conventional process of ground magnetic resonance detection comprises the steps of firstly collecting ground magnetic resonance signals by using an instrument, filtering noise contained in the signals by using an analog filter and a digital filter, and then carrying out inversion interpretation on the filtered signals to obtain underground aquifer information. Because the magnetic resonance signal is very weak, only in the order of a nano-volt, when the magnetic resonance signal is interfered by ambient noise, the magnetic resonance signal is often submerged in the noise. For spike noise, because the amplitude of the spike noise is far greater than the signal amplitude, the frequency distribution range is wide, and the data signal to noise ratio is seriously reduced, the effective elimination of the spike noise is very important for the subsequent inversion interpretation.
Patent CN106772646A discloses a "method for extracting a nuclear magnetic resonance signal on the ground", which uses a statistical method to determine whether spike noise exists, if so, removes the spike noise and replaces it with an interpolation result, if not, the measured data is kept unchanged. The method provides a thought for identifying and replacing the spike noise, but the spike noise identification process is complex, and due to the existence of power frequency and random noise, after the spike noise is removed, an interpolation result is used for conjecturing that an error exists in missing data, so that the accuracy of subsequent inversion is influenced.
Patent CN109100813A discloses a method for eliminating spike noise in ground nmr data based on collaborative filtering, which first judges whether spike noise exists in all measured data under an impulse moment according to the 3 σ law, and divides the measured data into two groups, i.e., spike noise-containing group and spike noise-free group. And then, respectively carrying out discrete cosine transform and Hadamard transform on the two groups of transform domain coefficients to obtain two groups of transform domain coefficients. And calculating a filter coefficient by using the transformation coefficient of the spike-free noise data, and filtering the coefficient containing the spike noise data. And finally, carrying out Hadamard and inverse discrete cosine transform on the filtered coefficient containing the peak noise data to eliminate the peak noise. The method does not delete or replace data containing spike noise time periods, does not introduce extra noise, but needs to repeatedly transmit the same pulse moment to acquire a plurality of groups of data, increases the measuring time of the instrument, and has large calculation amount and lower working efficiency.
The wavelet thresholding method is used in the paper "Despiking of magnetic resonance signals in time and wave domains" published by Stephan Costabel and Mike Muller-Petke at NEAR SURFACE chemistry 2014,12(2),185-197 to attenuate single spike noise present in magnetic resonance signals. The method can realize noise elimination processing of single data, improve detection efficiency, but the reconstructed signal is difficult to recover all characteristics of the magnetic resonance signal, and when the duration time of spike noise exceeds 10ms, signal components are obviously lost.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a ground magnetic resonance signal peak noise eliminating method combining an EMD algorithm and a WT algorithm to effectively eliminate a plurality of continuous peak noises appearing in a short time, the method has good applicability to low signal-to-noise ratio full-wave magnetic resonance signals interfered by the peak noises, an obvious noise reduction effect can be obtained by processing single acquisition signals, and the detection efficiency can be improved.
The present invention is achieved in such a way that,
a ground magnetic resonance signal spike noise elimination method combining an EMD algorithm and a WT algorithm comprises the following steps:
a. decomposing a ground magnetic resonance observation signal x (n) acquired by a nuclear magnetic resonance underground water detector by using an EMD algorithm to obtain a finite number of natural modal components c with frequencies from high to low1(n),…,ci(n) and a residual component resi(n), i is the number of modal components;
b. determining the individual natural modal components c1(n),…,ci(n) standard deviation, and screening out modal component c containing spike noise1(n),…,cj(n), j is the number of modal components containing spike noise;
c. adopting WT algorithm to screen out modal component c containing spike noise1(n),…,cj(n) respectively carrying out threshold processing, and only retaining spike noise components in each modal component to obtain spike noise modal components sp1(n),…,spj(n)；
d. The obtained peak noise modal component sp1(n),…,spj(n) accumulating and summing to obtain spike noise sp (n);
e. by means of difference operation, the peak noise sp (n) is eliminated from the magnetic resonance signal x (n) containing the peak noise, and a full-wave magnetic resonance signal s (n) without the interference of the peak noise is obtained, namely s (n) x (n) -sp (n).
Further, the EMD algorithm in step a includes the specific steps of:
a1) extracting all extreme points of the observation signals x (n);
a2) respectively fitting an upper envelope line and a lower envelope line to all the maximum values and the minimum values by utilizing a cubic spline interpolation principle;
a3) calculating the mean value m of the upper envelope and the lower envelope1(n) calculating a residual component h1(n)＝x(n)-m1(n), judgment of h1(n) whether the condition of the natural modal component is satisfied, and if so, recording h1(n) is the first natural mode component, the highest frequency component c of the observed signal x (n)1(n), otherwise let x (n) h1(n) looping the first two steps until a first eigenmode component is obtained;
a4) calculating residual error component res1(n)＝x(n)-c1(n) dividing residual component res1(n) repeating steps a1) through a3 as the original signal x (n) until the residual component resiStopping circulation when (n) is a monotonous function or a constant to obtain i natural modal components c1(n),…,ci(n) and a residual component resi(n), then: x (n) ═ c1(n)+…+ci(n)+resi(n)。
Further, the WT algorithm in step c is to process the modal component containing spike noise, and only retain spike noise components, and the specific steps are as follows:
c1) for modal component c containing spike noise1(n) performing wavelet decomposition to extract detail coefficients d of each layer1,…,dkAnd the last layer approximation coefficient akK is the number of wavelet decomposition layers;
c2) calculating the mean square error sigma of each layer coefficientk；
c3) The mean square error sigma of each layer coefficientkThreshold lambda is taken into accountk，
Wherein λkIs a real number, K is a constant, NkIs the length of the k-th layer detail coefficient;
c4) approximating the last layer of wavelet with coefficient akSetting zero, performing hard threshold processing on detail coefficients of each layer, setting the detail coefficients smaller than the obtained threshold to zero, and keeping the detail coefficients larger than the threshold unchanged;
c5) and reconstructing a peak noise modal component sp by using the last layer of wavelet approximation coefficient and each layer of detail coefficient after threshold processing1(n)；
c6) According to the processing methods from the step c1) to the step c5), sequentially carrying out the step c2(n),…,cj(n) processing to obtain peak noise modal component sp1(n),…,spj(n)。
Compared with the prior art, the invention has the beneficial effects that: according to the method, wavelet threshold processing is carried out on components containing the peak noise to reconstruct the peak noise component, then the peak noise is obtained through component summation, and the denoised signals are obtained through difference operation. Meanwhile, the method can effectively remove a plurality of continuous peak noises in a short time only by collecting signals once, does not need to repeatedly transmit the same pulse moment to collect a plurality of groups of signals, and greatly improves the detection efficiency.
Drawings
FIG. 1 is a flow chart of a method for eliminating noise of a spike of a ground magnetic resonance signal by combining EMD and WT algorithms;
FIG. 2 is a time domain plot of a magnetic resonance observation signal containing spike noise;
FIG. 3 shows the natural modal component and residual component of the peak noise-containing MR observation signal after EMD;
FIG. 4 is a diagram of a natural modal component containing spike noise;
FIG. 5 is a diagram of the eigenmode components of spike noise after wavelet thresholding;
figure 6 is a time domain plot of the magnetic resonance signal after spike noise cancellation.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1, a method for eliminating noise of a spike of a ground magnetic resonance signal by combining EMD and WT algorithms includes the following steps:
a. by using EThe MD algorithm decomposes a ground full-wave magnetic resonance observation signal x (n) acquired by the nuclear magnetic resonance underground water detector to obtain a finite number of inherent modal components c with frequencies from high to low1(n),…,ci(n) and a residual component resi(n), i is the number of modal components;
b. determining the individual natural modal components c1(n),…,ci(n) standard deviation, and screening out modal component c containing spike noise1(n),…,cj(n), j is the number of modal components containing spike noise;
c. according to the principle of WT algorithm, the screened modal component c containing spike noise is processed1(n),…,cj(n) respectively carrying out threshold processing, and only retaining spike noise components in each modal component to obtain spike noise modal components sp1(n),…,spj(n)；
d. The obtained peak noise modal component sp1(n),…,spj(n) accumulating and summing to obtain spike noise sp (n);
e. by means of difference operation, the peak noise sp (n) is eliminated from the magnetic resonance signal x (n) containing the peak noise, and a full-wave magnetic resonance signal s (n) without the interference of the peak noise is obtained, namely s (n) x (n) -sp (n).
The EMD algorithm in the step a comprises the following specific steps:
a1) extracting all extreme points of the observation signals x (n);
a2) respectively fitting an upper envelope line and a lower envelope line to all the maximum values and the minimum values by utilizing a cubic spline interpolation principle;
a3) calculating the average value m of the upper envelope and the lower envelope1(n) calculating a residual component h1(n)＝x(n)-m1(n), judgment of h1(n) whether the condition of the natural modal component is satisfied, and if so, recording h1(n) is the first natural modal component, i.e. the highest frequency component c of the signal x (n)1(n), otherwise let x (n) h1(n) looping the first two steps until a first eigenmode component is obtained;
a4) calculating residual error component res1(n)＝x(n)-c1(n) dividing the residual res1(n) as the original signal x (n), repeating steps 1) to 3) until the residual component resiStopping circulation when (n) is a monotonous function or a constant to obtain i natural modal components c1(n),…,ci(n) and a residual component resi(n) that is
x(n)＝c1(n)+…+ci(n)+resi(n)；
In the step c, the WT algorithm performs threshold processing on the modal component containing spike noise, and only retains spike noise components, and specifically includes the steps of:
c1) for modal component c containing spike noise1(n) performing wavelet decomposition to extract detail coefficients d of each layer1,…,dkAnd the last layer approximation coefficient akK is the number of wavelet decomposition layers;
c2) calculating the mean square error sigma of each layer coefficientk；
c3) The mean square error sigma of each layer coefficientkThreshold lambda is taken into accountk，
Wherein λkIs a real number, K is a constant, NkIs the length of the k-th layer detail coefficient;
4) approximating the last layer of wavelet with coefficient akAnd setting zero, and carrying out hard threshold processing on detail coefficients of each layer, namely setting the detail coefficients smaller than the threshold to zero, and keeping the detail coefficients larger than the threshold unchanged.
5) And reconstructing a peak noise modal component sp by using the last layer of wavelet approximation coefficient and each layer of detail coefficient after threshold processing1(n)；
6) According to the treatment methods from step 1) to step 5), sequentially carrying out the treatment on the C2(n),…,cj(n) processing to obtain peak noise modal component sp1(n),…,spj(n)。
Example (b):
detailed description of the ground magnetic resonance signal spike noise elimination method combining EMD and WT algorithms:
1) decomposing a ground full-wave magnetic resonance observation signal x (n) which is acquired by a nuclear magnetic resonance underground water detector and contains spike noise interference by utilizing an EMD algorithm, wherein the observation signal contains six spike noises which continuously appear in a short time, and the figure is shown in figure 2;
2) extracting all extreme points of the observation signals x (n);
3) respectively fitting an upper envelope line and a lower envelope line to all the maximum values and the minimum values by utilizing a cubic spline interpolation principle;
4) calculating the average value m of the upper envelope and the lower envelope1(n) calculating a residual component h1(n)＝x(n)-m1(n), judgment of h1(n) whether the condition of the natural modal component is satisfied, and if so, recording h1(n) is the first natural modal component, i.e. the highest frequency component c of the signal x (n)1(n), otherwise let x (n) h1(n) looping the first two steps until a first eigenmode component is obtained;
5) calculating residual error component res1(n)＝x(n)-c1(n) dividing the residual res1(n) as the original signal x (n), repeating steps 2) to 4) until the residual component resiStopping the cycle when (n) is a monotonic function or a constant to obtain 13 natural modal components c1(n),…,c13(n) and 1 residual component res13(n), i.e. x (n) ═ c1(n)+…+c13(n)+res13(n), the natural modal component and the residual component are shown in FIG. 3, and the left side is c from top to bottom1(n),…,c7(n) on the right side, from top to bottom, is c8(n),…,c13(n) and residual component res13(n); (results of the figure are explained)
6) And obtaining each natural mode component c1(n),…,c13(n) standard deviation, and 7 modal components c containing spike noise are screened out1(n),…,c7(n), see FIG. 4, is c from top to bottom1(n),…,c7(n)；
7) For modal component c containing spike noise1(n) performing wavelet decomposition to extract detail coefficients d of each layer1,…,dkAnd the last layer approximation coefficient ak；
8) Calculating the mean square error sigma of each layer coefficientk；
9) The mean square error sigma of each layer coefficientkThreshold lambda is taken into accountk，
Wherein λkIs a real number, K is a constant, NkIs the length of the k-th layer detail coefficient;
10) approximating the last layer of wavelet with coefficient akAnd setting zero, and carrying out hard threshold processing on detail coefficients of each layer, namely setting the detail coefficients smaller than the threshold to zero, and keeping the detail coefficients larger than the threshold unchanged.
11) And reconstructing a peak noise modal component sp by using the last layer of wavelet approximation coefficient and each layer of detail coefficient after threshold processing1(n)；
12) According to the processing methods from step 7) to step 11), sequentially carrying out the treatment on the C2(n),…,c7(n) processing to obtain peak noise modal component sp1(n),…,sp7(n), see FIG. 5; from top to bottom in order of sp1(n),…,sp7(n)；
13) And obtaining a peak noise modal component sp1(n),…,sp7(n) accumulating and summing to obtain spike noise sp (n);
14) and eliminating spike noise sp (n) from the magnetic resonance signal x (n) containing the spike noise by a difference operation to obtain a full-wave magnetic resonance signal s (n) without the spike noise interference, namely s (n) x (n) -sp (n), as shown in fig. 6, wherein a gray line represents a signal containing the spike noise, and a black line represents a signal after noise elimination.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (2)
1. A ground magnetic resonance signal spike noise elimination method combining an EMD algorithm and a WT algorithm is characterized by comprising the following steps:
a. decomposing a ground magnetic resonance observation signal x (n) acquired by a nuclear magnetic resonance underground water detector by using an EMD algorithm to obtain a finite number of natural modal components c with frequencies from high to low1(n),…,ci(n) and a residual component resi(n), i is the number of modal components;
b. determining the individual natural modal components c1(n),…,ci(n) standard deviation, and screening out modal component c containing spike noise1(n),…,cj(n), j is the number of modal components containing spike noise;
c. adopting WT algorithm to screen out modal component c containing spike noise1(n),…,cj(n) respectively carrying out threshold processing, and only retaining spike noise components in each modal component to obtain spike noise modal components sp1(n),…,spj(n)；
d. The obtained peak noise modal component sp1(n),…,spj(n) accumulating and summing to obtain spike noise sp (n);
e. removing spike noise sp (n) from the magnetic resonance signal x (n) containing the spike noise by a difference operation to obtain a full-wave magnetic resonance signal s (n) without spike noise interference, namely s (n) x (n) -sp (n);
the WT algorithm in step c is to process the modal component containing spike noise, and only retain spike noise components, and the specific steps are as follows:
c1) for modal component c containing spike noise1(n) performing wavelet decomposition to extract detail coefficients d of each layer1,…,dkAnd the last layer approximation coefficient akK is the number of wavelet decomposition layers;
c2) calculating the mean square error sigma of each layer coefficientk；
c3) The mean square error sigma of each layer coefficientkThreshold lambda is taken into accountk，
Wherein λkIs a real number, K is a constant, NkIs the firstLength of k-layer detail coefficients;
c4) approximating the last layer of wavelet with coefficient akSetting zero, performing hard threshold processing on detail coefficients of each layer, setting the detail coefficients smaller than the obtained threshold to zero, and keeping the detail coefficients larger than the threshold unchanged;
c5) and reconstructing a peak noise modal component sp by using the last layer of wavelet approximation coefficient and each layer of detail coefficient after threshold processing1(n)；
c6) According to the processing methods from the step c1) to the step c5), sequentially carrying out the step c2(n),…,cj(n) processing to obtain peak noise modal component sp1(n),…,spj(n)；
The spike noise is a noise which appears continuously in a short time and has a much larger amplitude than the signal amplitude.
2. The method for ground magnetic resonance signal spike noise rejection combining EMD and WT algorithms of claim 1,
the EMD algorithm in the step a comprises the following specific steps:
a1) extracting all extreme points of the observation signals x (n);
a2) respectively fitting an upper envelope line and a lower envelope line to all the maximum values and the minimum values by utilizing a cubic spline interpolation principle;
a3) calculating the mean value m of the upper envelope and the lower envelope1(n) calculating a residual component h1(n)＝x(n)-m1(n), judgment of h1(n) whether the condition of the natural modal component is satisfied, and if so, recording h1(n) is the first natural mode component, the highest frequency component c of the observed signal x (n)1(n), otherwise let x (n) h1(n) looping the first two steps until a first eigenmode component is obtained;
a4) calculating residual error component res1(n)＝x(n)-c1(n) dividing residual component res1(n) repeating steps a1) through a3 as the original signal x (n) until the residual component resiStopping circulation when (n) is a monotonous function or a constant to obtain i natural modal components c1(n),…,ci(n) and a residual component resi(n), then:
x(n)＝c1(n)+…+ci(n)+resi(n)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010401183.3A CN111650654B (en) | 2020-05-13 | 2020-05-13 | Ground magnetic resonance signal peak noise elimination method combining EMD and WT algorithms |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010401183.3A CN111650654B (en) | 2020-05-13 | 2020-05-13 | Ground magnetic resonance signal peak noise elimination method combining EMD and WT algorithms |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111650654A CN111650654A (en) | 2020-09-11 |
| CN111650654B true CN111650654B (en) | 2022-03-22 |
Family
ID=72342071
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010401183.3A Active CN111650654B (en) | 2020-05-13 | 2020-05-13 | Ground magnetic resonance signal peak noise elimination method combining EMD and WT algorithms |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111650654B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112434634B (en) * | 2020-12-02 | 2022-10-11 | 青岛理工大学 | Method and system for rapidly eliminating civil engineering structure health monitoring signal peak |
| CN113376698B (en) * | 2021-06-23 | 2022-09-27 | 吉林大学 | Low-missing-detection seismic data peak noise detection and suppression method |
| CN115017933B (en) * | 2022-03-11 | 2024-04-05 | 吉林大学 | Nuclear magnetic signal peak noise suppression method based on quasi-periodic wavelet coefficient recovery |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107783200A (en) * | 2017-11-21 | 2018-03-09 | 吉林大学 | Joint EMD and TFPF algorithms a kind of all-wave magnetic resonance signal random noise method for reducing |
| CN108051857A (en) * | 2017-12-05 | 2018-05-18 | 中国石油大学(华东) | A kind of local similarity wavelet threshold denoising method based on empirical mode decomposition |
| WO2018090467A1 (en) * | 2016-11-17 | 2018-05-24 | 丹阳华神电器有限公司 | Fuzzy entropy based noisy signal processing method and iterative singular spectrum soft thresholding denoising method |
| 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 |
| CN111007569A (en) * | 2019-12-11 | 2020-04-14 | 中国人民解放军96901部队23分队 | Sample entropy threshold microseism signal noise reduction method integrated with empirical mode decomposition |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109633761B (en) * | 2018-12-13 | 2021-05-28 | 吉林大学 | Magnetic resonance signal power frequency noise reduction method based on wavelet transformation modulus maximum value method |
-
2020
- 2020-05-13 CN CN202010401183.3A patent/CN111650654B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018090467A1 (en) * | 2016-11-17 | 2018-05-24 | 丹阳华神电器有限公司 | Fuzzy entropy based noisy signal processing method and iterative singular spectrum soft thresholding denoising method |
| CN107783200A (en) * | 2017-11-21 | 2018-03-09 | 吉林大学 | Joint EMD and TFPF algorithms a kind of all-wave magnetic resonance signal random noise method for reducing |
| CN108051857A (en) * | 2017-12-05 | 2018-05-18 | 中国石油大学(华东) | A kind of local similarity wavelet threshold denoising method based on empirical mode decomposition |
| 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 |
| CN111007569A (en) * | 2019-12-11 | 2020-04-14 | 中国人民解放军96901部队23分队 | Sample entropy threshold microseism signal noise reduction method integrated with empirical mode decomposition |
Non-Patent Citations (2)
| Title |
|---|
| 基于EMD和小波变换的核磁测井回波信号去噪研究;孙灵川;《中国优秀硕士学位论文全文数据库 基础科学辑》;20150228(第02期);创新点摘要及正文第9-16、21、25-30、41-48页 * |
| 基于EMD小波阈值滤波的地磁数据去噪分析;张敏 等;《地震地磁观测与研究》;20120630;第33卷(第3期);第171-175页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111650654A (en) | 2020-09-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111650654B (en) | Ground magnetic resonance signal peak noise elimination method combining EMD and WT algorithms | |
| Hao et al. | A novel multiscale nonlinear thresholding method for ultrasonic speckle suppressing | |
| CN105205788B (en) | A kind of denoising method for high-throughput gene sequencing image | |
| CN109633761B (en) | Magnetic resonance signal power frequency noise reduction method based on wavelet transformation modulus maximum value method | |
| CN111680785B (en) | ECG signal processing method based on sparse characteristic and antagonistic neural network combination | |
| CN106771905B (en) | A kind of DISCHARGE PULSES EXTRACTION method suitable for high-frequency current Partial Discharge Detection | |
| CN110646851B (en) | Adaptive threshold seismic random noise suppression method based on Shearlet transformation | |
| CN110261910A (en) | Seismic data surface wave minimizing technology based on adaptive sparse S-transformation | |
| CN111046791A (en) | Current signal filtering and denoising method based on generalized S transform containing variable factors | |
| CN110515063A (en) | Underwater acoustic signal processing method and apparatus based on the steady wavelet transform of iteration | |
| CN109212608B (en) | Borehole microseismic signal antinoise method based on 3D shearlet transformation | |
| Placidi et al. | Post-processing noise removal algorithm for magnetic resonance imaging based on edge detection and wavelet analysis | |
| Ou et al. | Frame-based time-scale filters for underwater acoustic noise reduction | |
| CN110032968A (en) | Denoising method based on dual-tree complex wavelet and adaptive semi-soft threshold model | |
| CN115657140A (en) | Magnetic anomaly detection method and system based on structured Hankel total variation regularization | |
| CN114358040A (en) | Disturbing signal denoising method based on adaptive estimation threshold method | |
| CN113505688A (en) | Improved self-adaptive wavelet threshold signal denoising method | |
| Kahoo et al. | Random noise suppression from seismic data using time-frequency peak filtering | |
| Chen et al. | Research on sonar image denoising method based on fixed water area noise model | |
| Ferzo et al. | Digital image denoising techniques in wavelet domain with another filter: A review | |
| CN110575145A (en) | Enhanced Gaussian redundant dictionary pulse signal denoising method | |
| CN112363218A (en) | 3D-Shearlet transformation and scale self-adaption based seismic random noise suppression method | |
| CN114859419A (en) | Doppler high-frequency signal noise reduction method and system | |
| CN112711074B (en) | Denoising method and device for seismic first arrival wave | |
| CN115730190A (en) | Denoising method and system for partial discharge signal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |