CN109884080A - FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm - Google Patents
FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm Download PDFInfo
- Publication number
- CN109884080A CN109884080A CN201910166576.8A CN201910166576A CN109884080A CN 109884080 A CN109884080 A CN 109884080A CN 201910166576 A CN201910166576 A CN 201910166576A CN 109884080 A CN109884080 A CN 109884080A
- Authority
- CN
- China
- Prior art keywords
- peak
- central wavelength
- spectrum
- detection algorithm
- fbg
- 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
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The present invention provides a kind of FBG central wavelength Peak Search Method of adaptive multimodal detection algorithm, it combines complete non-recursive variation mode decomposition method VMD-DWT to carry out wavelet threshold denoising to FBG signal the following steps are included: step 1 using improved wavelet threshold Denoising Algorithm；Step 2, using multi-peak central wavelength detection algorithm, Hilbert transformation is carried out to denoising spectrum signal, then confirms the pre-determined bit of target peak in multimodal spectrum, and carry out peak region segmentation；Step 3, the central wavelength at multimodal spectral target peak, and output center wavelength are detected by region centroid method.The present invention provides a kind of fiber bragg grating (FBG) central wavelength Peak Search Method based on adaptive multimodal detection algorithm, the central wavelength of FGB reflectance spectrum can fast, accurately be found, by extracting the variation relation with structural damage feature of central wavelength, the monitoring to the Crack Damage state of aluminum alloy plate materials can be realized.
Description
Technical field:
The invention belongs to structural health monitoring technology field more particularly to a kind of light based on adaptive multimodal detection algorithm
Fine Bragg grating (FBG) central wavelength Peak Search Method.
Background technique:
Monitoring structural health conditions (Structural Health Monitoring) technology be advanced sensor is embedded to or
The mode of adhesive surface is arranged in monitored inside configuration or outside, for perceiving the configuration state in monitored region (as tied
Stress, strain, crackle of structure etc.) and external environment (such as temperature, pressure) variation.Structural health monitoring technology has to knot
The health status of structure carries out the advantages of monitoring in real time and finding structural damage in time, by being detected and being positioned to damage,
Total is set to be in monitored state always.Sensor more commonly used at present has: fiber-optic grating sensor, piezoelectric sensing
Device, smart coat sensor etc..Wherein, it is supervised in real time the stress/strain field that fiber-optic grating sensor can paste region to its cloth
It surveys, and has many advantages, such as electromagnetism interference, anticorrosive, high sensitivity, light-weight, corrosion-resistant, have in terms of monitoring structural health conditions
Have broad application prospects.
Since aircraft equipment is for assemble or the requirement of specific function, onboard drilling will lead to partial region stress collection
In, it then will lead to the further different degrees of damage of structure under external load effect.How skill monitored by FBG sensor
Art is monitored structural crack, is one long-term the problem of attracting attention.In addition, external loading can lead to along sensor grating
Uniformly or non-uniformly strain field distribution.When grating is under heterogeneous strain, reflectance spectrum can be distorted (as reflected
Spectrum broadens, reflectance spectrum generates multimodal and center wavelength shift etc.).In order to monitor the faulted condition of this complexity, many is ground
Study carefully and passed through experiment and simulated reflectance spectrum of the embedded FBG sensor under composite laminated plate damage propatagtion, but very
The relationship of rare research aluminium alloy plate transverse direction hole edge crack Propagation and FBG sensing characteristics.
And traditional peak search algorithm, such as maximum and extreme value algorithm, first derivative and threshold method have many offices
It is sex-limited, if noiseproof feature is poor, computational accuracy is low etc..Peak value searching method based on curve matching includes Gauss curve fitting, multinomial
Fitting, 3 peak detections and centroid algorithm Peak detection accuracy with higher, but its performance is influenced by spectrum types, it is special
It is not for deforming asymmetry spectrum.
Summary of the invention
To solve the above problems, the present invention provides a kind of fiber bragg grating based on adaptive multimodal detection algorithm
(FBG) central wavelength Peak Search Method can fast, accurately find the central wavelength of FGB reflectance spectrum, pass through cardiac wave in extracting
The monitoring to the Crack Damage state of aluminum alloy plate materials can be realized in the long variation relation with structural damage feature.
A kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm of the invention comprising following step
It is rapid:
Step 1, complete non-recursive variation mode decomposition method VMD-DWT is combined using improved wavelet threshold Denoising Algorithm
Wavelet threshold denoising is carried out to FBG signal；
Step 2, using multi-peak central wavelength detection algorithm, Hilbert transformation is carried out to denoising spectrum signal, then really
Recognize the pre-determined bit of target peak in multimodal spectrum, and carries out peak region segmentation；
Step 3, the central wavelength at multimodal spectral target peak, and output center wavelength are detected by region centroid method；
Wherein, complete non-recursive variation Mode Decomposition method VMD described in step 1 can be mentioned from input signal simultaneously
It takes and centre frequency wkRelevant mode, the mode are known as band limit intrinsic mode function BLIMFukBasic model, correspond to letter
Number sub- energy；When decomposition number K is predefined, the Energy distribution under every kind of basic model is also determined.
Further, step 1 includes:
Step 11,6 and optimum balance are set by the best decomposition number K of complete non-recursive variation mode decomposition method
Parameter alpha is equal to 200, then is decomposed into given K by complete non-recursive variation mode VMD with the input signal that length is N
Band limit intrinsic mode function BLIMFuk；
Step 12, threshold transformation is carried out to each BLIMF, part is carried out to the part high-order BLIMF that noise seriously destroys
It rejects；And 6 are set as wavelet basis, and by decomposition rank using DWT sym5；
Step 13, after carrying out denoising to each scale respectively, all threshold denoising blocks are added, rebuild denoising letter
Number.
Further, the wavelet threshold in improved wavelet threshold Denoising Algorithm are as follows:
Wherein, diIndicate the i-th threshold of BLIMFs, the standard deviation of noise is by component intermediate valueEstimation；
Description corresponds to the threshold of each mode of the variance of noise energy
Value Ti；Indicate k-th of energy of BLIMFs；Wherein, E1It is the energy of first BLIMFs
Amount, and iterative process parameter beta is shifted, ρ specifically gives as 0.719 and 2.01.
Further, multi-peak central wavelength detection algorithm is divided into two subtasks in step 2, and as multimodal segmentation is in
Heart wavelength detecting, wherein multimodal segmentation includes separation reflectance spectrum and finds multi-peak approximate location；Central wavelength is detected from every
Central wavelength is extracted in a divided reflectance spectrum.
Further, centroid algorithm generates the point for corresponding to spectrum geometry mass center, the corresponding wavelength of point in step 3
The as central wavelength at the peak.
The utility model has the advantages that adaptive multimodal algorithm proposed by the invention can efficiently extract the damage of center wavelength shift
Feature.
Use the true center wavelength in centroid algorithm detection multimodal.And super-Gaussian model is difficult to that modified function is selected to join
Number, Monte Carlo method are difficult to meet required precision since its is non-linear；Several optimization algorithms, such as genetic algorithm, adaptively
Neighborhood search, tree search, dynamic multigroup particle optimization algorithm handle multimodal test problems；But iteration takes a long time
Best solution just can be found and computation complexity is very high.Compared with these algorithms, centroid algorithm of the invention is being handled
Precision and robustness more with higher than direct peak value location algorithm when noise initial data.In addition, the calculation amount of centroid algorithm
Than in FBG reflectance spectrum Gauss or fitting of a polynomial it is high.
Detailed description of the invention
Fig. 1 is the FBG central wavelength Peak Search Method flow chart of the invention based on adaptive multimodal detection algorithm.
Specific embodiment
Fiber bragg grating (FBG) central wavelength Peak Search Method based on adaptive multimodal detection algorithm, external loading
It can lead to the uniformly or non-uniformly strain field distribution along sensing grating, reflectance spectrum caused to distort.For distortion spectrum,
There are certain defects for traditional peak-seeking algorithm.It is therefore desirable to propose that a kind of quick spectrum multimodal detection algorithm is accurately sought
Look for the central wavelength of FGB reflectance spectrum.
Based on this principle, the present invention proposes a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm,
Then the reflectance spectrum obtained using the FBG sensor that cloth is attached to aluminium alloy plate body structure surface uses multimodal proposed by the present invention
Detection algorithm, for detecting FBG reflectance spectrum central wavelength with the variation of crack propagation.It is thus determined that in FBG reflectance spectrum
Cardiac wave length is an important problem.By extracting the variation relation with structural damage feature of central wavelength, can be realized pair
The monitoring of the Crack Damage state of aluminum alloy plate materials.
As shown in Figure 1, the FBG central wavelength Peak Search Method of the invention based on adaptive multimodal detection algorithm includes following
Step:
Step 1, FBG signal is denoised using improved wavelet threshold Denoising Algorithm combination variation mode decomposition method.
FBG sensor-based system is by the shape of reflectance spectrum of the interference effect from external noise (such as electromagnetism and load).Reflected light
The distortion of spectrum is such as: spectrum widening, asymmetric, and the sharp peaks characteristics such as top fluctuation and secondary lobe limit Peak detection accuracy.This should
Top fluctuation is eliminated using effective smoothing denoising method.Therefore, FBG signal is denoised before peak detection.
The present invention using a kind of method for being combined with wavelet threshold denoising method of non-recursive variation mode decomposition completely come
Handle FBG signal.Wherein, completely non-recursive variation Mode Decomposition (VMD), can simultaneously from input signal extract and center
Frequency wkRelevant mode.The basic model is known as band limit intrinsic mode function (BLIMF) ukBasic model, correspond to signal
Sub- energy.When decomposition number K is predefined, the Energy distribution under every kind of basic model is determined.Then, it is widely used in
The inspiration of the translation-invariant feature of signal processing technology, the present invention propose that non-recursive variation Mode Decomposition combines improved completely
Wavelet threshold denoising technology (VMD-DWT).
Specific step 1 includes the following contents:
Step 11,6 are set by best decomposition number K and optimum balance parameter alpha is equal to 200.The input for being N with length
Signal is decomposed into K given band limit intrinsic mode function (BLIMF) u by complete non-recursive variation mode (VMD)k。
Step 12, threshold transformation, on the basis of literature research, the height that seriously destroys to noise are carried out to each BLIMF
The part rank BLIMF carries out local rejecting.Then, using DWT sym5 as wavelet basis, and 6 are set by decomposition rank.
When for FBG reflectance spectrum signal denoising, soft-threshold has better effect than hard -threshold.Then, in the present invention,
Soft-threshold is applied to all BLIMFs samples.In addition, extreme value needs in a smooth manner when the extreme value of sample is more than threshold value
Reduce the amount for being equal to threshold value.Consider each basic model ukIn Energy distribution influence when, improved wavelet threshold such as institute in formula 1
Show.
Here, diIndicate the i-th threshold of BLIMFs, the standard deviation of noise is by component intermediate valueEstimation.In addition, formula 2 describes the threshold value of each mode of the variance corresponding to noise energy
Ti。
Here, EiIt indicates i-th of energy of BLIMFs, and can be estimated by formula 2.
Here, E1It is the energy of first BLIMFs, and shifts iterative process parameter beta, ρ specifically gives as 0.719 He
2.01。
Step 13, after carrying out denoising to each scale respectively, all threshold denoising blocks are added, rebuild denoising letter
Number.
It is handled by above-mentioned improvement wavelet method, has effectively removed the noise for including in spectral signal, for multimodal detection behaviour
It lays a good foundation.The wavelet method that VMD changes effective filters out noise and retains details from partial stack signal.
Step 2, using multi-peak central wavelength detection algorithm, the denoising spectrum signal with Hilbert transformation is handled, really
Recognize the quantity and range of target peak in multimodal spectrum.Pre-determined bit and peak region segmentation including peak value.
Wherein, multi-peak central wavelength detection algorithm is for handling accurate multi-peak test problems, especially with regard to change
Shape reflectance spectrum.In innovatory algorithm, it is divided into two subtasks, as multimodal segmentation and central wavelength detection.Wherein, multimodal point
Steamed sandwich includes separation reflectance spectrum and finds multi-peak approximate location.Central wavelength detection is mentioned from each divided reflectance spectrum
Take central wavelength.
Specifically, step 2 includes the following contents:
Step 21, the Hilbert transformation in time domain is practical continuous spectra signal x (t) and (π t)-1Between convolution integral.
Then the Hilbert transform of x (t) is expressed as equation 4 by us.
According to one of Hilbert transform relationship, only phase spectrum is in x(t)Be different between x (t), and amplitude spectrum and
Energy spectrum is identical.
Step 22, threshold value is set.It first determines the Threshold range at unimodal place, and is defined as subsequent multimodal peak-seeking
Threshold value.Since the grid points on different fiber-optic grating sensors are different, threshold range is also different.Generally using observation
Method determines the threshold range of particular fiber grating sensor.
Step 23, spectrum is separated.Left and right extreme point is only poor to be greater than theoretical initial threshold.The peak of as one spectrum.It utilizes
This method determines the quantity and range at the peak of spectrum.
Step 24,3db threshold value bandwidth is chosen.Compared with traditional normal threshold peak detection method, set forth herein 3db from
Dynamic peak detection algorithm overcomes due to there are secondary lobe or peak distortion, caused by low value peak value loss detection and false peaks inspection
The limitation of survey.
Step 3, the central wavelength of multimodal is detected by centroid method.Centroid algorithm generates one and corresponds to spectrum geometry mass center
Point, which is the central wavelength at the peak.
It is calculated by formula 5, wherein N is the size of spectrum point vector, λiIt is i-th wavelength, IiIt is that i-th of point reflection rate is strong
Degree.
In addition, it should be understood that although this specification is described in terms of embodiments, but not each embodiment is only wrapped
Containing an independent technical solution, this description of the specification is merely for the sake of clarity, and those skilled in the art should
It considers the specification as a whole, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art
The other embodiments being understood that.
Claims (7)
1. a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm, which is characterized in that including following step
It is rapid:
Step 1, complete VMD-DWT pairs of method of non-recursive variation mode decomposition is combined using improved wavelet threshold Denoising Algorithm
FBG signal carries out wavelet threshold denoising；
Step 2, using multi-peak central wavelength detection algorithm, Hilbert transformation is carried out to denoising spectrum signal, is then confirmed more
The pre-determined bit of target peak in peak spectrum, and carry out peak region segmentation；
Step 3, the central wavelength at multimodal spectral target peak, and output center wavelength are detected by region centroid method；
Wherein, complete non-recursive variation Mode Decomposition method VMD described in step 1, can simultaneously from input signal extract with
Centre frequency wkRelevant mode, the mode are known as band limit intrinsic mode function BLIMFukBasic model, corresponding to signal
Sub- energy；When decomposition number K is predefined, the Energy distribution under every kind of basic model is also determined.
2. a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm as described in claim 1, feature
It is, step 1 includes:
Step 11,6 and optimum balance parameter are set by the best decomposition number K of complete non-recursive variation mode decomposition method
α is equal to 200, then K given band limit is decomposed by complete non-recursive variation mode VMD with the input signal that length is N
Intrinsic mode function BLIMFuk；
Step 12, threshold transformation is carried out to each BLIMF, local rejecting is carried out to the part high-order BLIMF that noise seriously destroys；
And 6 are set as wavelet basis, and by decomposition rank using DWT sym5；
Step 13, after carrying out denoising to each scale respectively, all threshold denoising blocks are added, rebuild denoised signal.
3. a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm as described in claim 1, feature
It is, the wavelet threshold in improved wavelet threshold Denoising Algorithm are as follows:
Wherein, diIndicate the i-th threshold of BLIMFs, the standard deviation of noise is by component intermediate valueEstimation；
Description corresponds to the threshold value of each mode of the variance of noise energy
Ti；Indicate k-th of energy of BLIMFs；Wherein, E1It is the energy of first BLIMFs,
And iterative process parameter beta is shifted, ρ specifically gives as 0.719 and 2.01.
4. a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm as described in claim 1, feature
It is,
Multi-peak central wavelength detection algorithm is divided into two subtasks, as multimodal segmentation and central wavelength detection in step 2,
In, multimodal segmentation includes separation reflectance spectrum and finds multi-peak approximate location；Central wavelength is detected from each divided anti-
It penetrates and extracts central wavelength in spectrum.
5. a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm as claimed in claim 4, feature
It is, step 2 includes with sub-step:
Step 21, the Hilbert transformation in time domain is practical continuous spectra signal x (t) and (π t)-1Between convolution integral, it is as follows
Formula indicates
Step 22, threshold value is set
It first determines the Threshold range at unimodal place, and is defined as the threshold value of subsequent multimodal peak-seeking；
Step 23, spectrum is separated
The difference of left and right extreme point is greater than theoretical initial threshold, the peak of as one spectrum, with the quantity at this peak for determining spectrum and
Range；
Step 24,3db threshold value bandwidth is chosen.
6. a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm as described in claim 1, feature
It is,
Centroid algorithm generates the point for corresponding to spectrum geometry mass center in step 3, which is in the peak
Cardiac wave is long.
7. a kind of FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm as claimed in claim 6, feature
It is, the central wavelength at the peak is
Wherein N is the size of spectrum point vector, λiIt is i-th wavelength, IiIt is i-th of point reflection rate intensity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910166576.8A CN109884080A (en) | 2019-03-06 | 2019-03-06 | FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910166576.8A CN109884080A (en) | 2019-03-06 | 2019-03-06 | FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109884080A true CN109884080A (en) | 2019-06-14 |
Family
ID=66930858
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910166576.8A Pending CN109884080A (en) | 2019-03-06 | 2019-03-06 | FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109884080A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110542441A (en) * | 2019-10-10 | 2019-12-06 | 华北电力大学（保定） | Signal demodulation method of optical fiber Bragg grating sensing system |
| CN111044471A (en) * | 2019-12-31 | 2020-04-21 | 中国电子科技集团公司信息科学研究院 | Crack damage monitoring method based on damage characteristic parameter extraction algorithm |
| CN111208142A (en) * | 2019-08-01 | 2020-05-29 | 北京航空航天大学 | Crack damage quantitative detection method based on dynamic time warping correlation characteristics |
| CN113221692A (en) * | 2021-04-29 | 2021-08-06 | 长春工业大学 | Continuous variational modal decomposition DWT denoising method for optical fiber sensing |
| CN113358239A (en) * | 2021-05-24 | 2021-09-07 | 长春工业大学 | FBG-based wavelength feature identification method |
| CN113407908A (en) * | 2021-07-14 | 2021-09-17 | 北京华大九天科技股份有限公司 | Method for vector fitting in multi-peak frequency spectrum |
| CN113639891A (en) * | 2020-09-03 | 2021-11-12 | 深圳阿珂法先进科技有限公司 | High-speed optical fiber temperature sensing demodulation method based on equivalent wavelength |
| CN113839711A (en) * | 2021-09-10 | 2021-12-24 | 深圳技术大学 | Peak detection method based on dynamic threshold distance centroid algorithm |
| CN116907556A (en) * | 2023-09-11 | 2023-10-20 | 武汉理工大学 | Distributed optical fiber sensing multi-feature hybrid demodulation system and method |
-
2019
- 2019-03-06 CN CN201910166576.8A patent/CN109884080A/en active Pending
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111208142A (en) * | 2019-08-01 | 2020-05-29 | 北京航空航天大学 | Crack damage quantitative detection method based on dynamic time warping correlation characteristics |
| CN110542441A (en) * | 2019-10-10 | 2019-12-06 | 华北电力大学（保定） | Signal demodulation method of optical fiber Bragg grating sensing system |
| CN110542441B (en) * | 2019-10-10 | 2021-08-27 | 华北电力大学（保定） | Signal demodulation method of optical fiber Bragg grating sensing system |
| CN111044471A (en) * | 2019-12-31 | 2020-04-21 | 中国电子科技集团公司信息科学研究院 | Crack damage monitoring method based on damage characteristic parameter extraction algorithm |
| CN113639891A (en) * | 2020-09-03 | 2021-11-12 | 深圳阿珂法先进科技有限公司 | High-speed optical fiber temperature sensing demodulation method based on equivalent wavelength |
| CN113221692A (en) * | 2021-04-29 | 2021-08-06 | 长春工业大学 | Continuous variational modal decomposition DWT denoising method for optical fiber sensing |
| CN113358239A (en) * | 2021-05-24 | 2021-09-07 | 长春工业大学 | FBG-based wavelength feature identification method |
| CN113407908A (en) * | 2021-07-14 | 2021-09-17 | 北京华大九天科技股份有限公司 | Method for vector fitting in multi-peak frequency spectrum |
| CN113839711A (en) * | 2021-09-10 | 2021-12-24 | 深圳技术大学 | Peak detection method based on dynamic threshold distance centroid algorithm |
| CN113839711B (en) * | 2021-09-10 | 2022-08-02 | 深圳技术大学 | Peak detection method based on dynamic threshold distance centroid algorithm |
| CN116907556A (en) * | 2023-09-11 | 2023-10-20 | 武汉理工大学 | Distributed optical fiber sensing multi-feature hybrid demodulation system and method |
| CN116907556B (en) * | 2023-09-11 | 2024-04-16 | 武汉理工大学 | Distributed optical fiber sensing multi-feature hybrid demodulation system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109884080A (en) | FBG central wavelength Peak Search Method based on adaptive multimodal detection algorithm | |
| CN108955734B (en) | Cavity length demodulation method of optical fiber F-P temperature/pressure composite sensor | |
| Nardi et al. | Detection of low-velocity impact-induced delaminations in composite laminates using Auto-Regressive models | |
| CN106771905B (en) | A kind of DISCHARGE PULSES EXTRACTION method suitable for high-frequency current Partial Discharge Detection | |
| CN108225385A (en) | A kind of overlapping FBG transducing signal peak value localization methods | |
| CN107490397A (en) | High-accuracy self-adaptation filters the quick Peak Search Method of FBG spectrum | |
| CN108645620A (en) | A kind of Fault Diagnosis of Rolling Element Bearings method based on comentropy and Multiscale Morphological | |
| CN104408288A (en) | Multiple-stable-state stochastic resonance weak signal detection method based on wavelet and parameter compensation | |
| CN111181634B (en) | Distributed optical fiber vibration signal rapid positioning method | |
| CN109815849A (en) | Chaotic signal Denoising Algorithm based on singular value decomposition | |
| Potapov | Textures, fractals, scaling effects, and fractional operators as a basis of new methods for data processing and fractal radio system design | |
| CN113887398A (en) | GPR signal denoising method based on variational modal decomposition and singular spectrum analysis | |
| CN105181120A (en) | High-sensitivity transformer winding loosening determination method | |
| CN107966206A (en) | A kind of array signal denoising method of fiber ultrasonic sensor array | |
| Li et al. | Spectral-Spatial Depth-Based Framework for Hyperspectral Underwater Target Detection | |
| CN113158830A (en) | Residual gravity abnormal field separation method | |
| CN104573732A (en) | Target spectral matching method | |
| Meruane et al. | Impact identification using nonlinear dimensionality reduction and supervised learning | |
| CN114061848B (en) | Method for identifying leak hole of reinforced sealing structure of spacecraft | |
| Janarthan et al. | Damage detection in stiffened composite panels using Lamb wave | |
| CN110378294B (en) | Hyperspectral target detection method and system based on local energy constraint and feature fusion | |
| Shi et al. | Wind turbines fault diagnosis method under variable working conditions based on AMVMD and deep discrimination transfer learning network | |
| Cao et al. | Application of wavelet packet energy spectrum to extract the feature of the pulse signal | |
| Zhang et al. | Low-rank and periodic group sparse based signal denoising method for rolling bearing fault feature extraction | |
| Chen et al. | Fusarium wilt inspection for Phalaenopsis using uniform interval hyperspectral band selection techniques |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190614 |
|
| WD01 | Invention patent application deemed withdrawn after publication |