CN107727749A - A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm - Google Patents
A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm Download PDFInfo
- Publication number
- CN107727749A CN107727749A CN201710760358.8A CN201710760358A CN107727749A CN 107727749 A CN107727749 A CN 107727749A CN 201710760358 A CN201710760358 A CN 201710760358A CN 107727749 A CN107727749 A CN 107727749A
- Authority
- CN
- China
- Prior art keywords
- feature extraction
- extraction algorithm
- wavelet packet
- detection method
- mrow
- 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
- G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
Abstract
The invention discloses a kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm, including step are as follows：The surface wave signal of flaw echo is gathered, and carries out waveform interception；Gone using coif3 small echos dry；Defect characteristic is extracted using wavelet packet fusion feature extraction algorithm；Classification quantitative judge is carried out using sorting algorithm.The method of the present invention need not tangle the ultrasonic reflection modeling in complexity, have preferable versatility and practicality.
Description
Technical field
The present invention devises a kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm, belongs to lossless
Detection technique field.
Background technology
Non-Destructive Testing is using characteristics such as sound, light, magnetic and electricity, is not damaging or is not influenceing checked object performance
Under the premise of, detect in checked object and whether there is defect or inhomogeneities, provide size, position, property and quantity of defect etc.
Information, and then judge the general name of all technological means of state of the art residing for checked object, it is that industrial development is essential
Effective tool, to a certain extent reflect a national Industry Development Level, its importance obtained the world it is widely recognized that.I
Government of state also pays much attention to the research and application of non-destructive testing technology, and nationwide Non-Destructive Testing science has been set up in 1978
Tissue-Non-Destructive Testing branch of Chinese Mechanical Engineering Society.At the same time, country is planned by " 863 ", " 973 " plan and state
Multiple approach such as family's Natural Science Fund In The Light, support the development of China's non-destructive testing technology energetically.In various non-destructive testing technologies,
Ultrasound detection with the advantages that its cost is low, speed is fast, harmless, applied widely and accurate positioning to defect, into
For field of non destructive testing most popular method.
Waveform, amplitude and the spectrum distribution of ultrasonic reflection echo-signal carry the information of defect, non-imaging type ultrasound
Default kind identification method is detected, can be by extracting defect information feature, analysis characteristic parameter and defect class from detection signal
Corresponding relation between type, realize the type identification to defect.This method need not tangle the ultrasonic reflection modeling in complexity, tool
There are preferable versatility and practicality.Non-imaging type defects in ultrasonic testing type identification is the weight of online ultrasound detection quantitative analysis
Will basis, non-imaging type defects in ultrasonic testing kind identification method of the research with high reliability be online automatic for realizing
Ultrasound detection and evaluation, the efficiency of raising Ultrasonic NDT and quality all have great importance.
The content of the invention
It is an object of the invention to for online automatic ultrasonic detection with evaluation, improve Ultrasonic NDT efficiency and
A kind of quality, it is proposed that ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm.This method is first with laser
The surface wave signal of Experimental Ultrasonic system acquisition flaw echo, then using coif3 Wavelet Denoising Methods, then it is special using wavelet packet fusion
Extraction algorithm extraction defect characteristic is levied, finally carries out classification quantitative judge using sorting algorithm.
To reach above-mentioned purpose, the technical solution adopted by the present invention is as follows：
A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm, including step are as follows：
Step 1：Using the surface wave signal of laser-ultrasound experimental system collection flaw echo, and carry out waveform interception；
Step 2：Gone using coif3 small echos dry；
Step 3：Defect characteristic is extracted using wavelet packet fusion feature extraction algorithm；
Step 4：Classification quantitative judge is carried out using sorting algorithm.
Preferably, the step 2 specifically includes：Using five layers of wavelet soft-threshold denoising.
Preferably, the step 2 specifically also includes：Five layers of wavelet soft-threshold use rigrsure threshold values, and the threshold values is
Using the adaptive threshold selection based on Stein unbiaseds/possibility predication principle, a threshold values λ is given, the likelihood for obtaining it is estimated
Meter, then non-likelihood λ is minimized, obtain selected threshold values.
Preferably, the feature extraction algorithm in the step 3 specifically includes：
31) sub-band coefficients are obtained using wavelet package transforms and calculates the Energy distribution of each node, choose effective node；
X=[x00,x01,x02,...,xij,xij+1,...,xmn]
Vectorial X includes all wavelet packet coefficients, wherein, i represents nodes, and j represents the coefficient number in node, m i
Sum, n be j sum；
32) correlation analysis is carried out to effective node, data is mapped to higher dimensional space by kernel function, further according to correlation
Property yojan and conversion carried out to this class sample obtain feature vector, Xnew：
Xnew=[α1,...,αm][k(X1,Xnew),...,k(Xm,Xnew)]T
Wherein, α is the characteristic vector of nuclear matrix, and k is nuclear matrix；
33) effective node is analyzed, obtains the energy Spectral structure of each node using frequency spectrum algorithm, obtain characteristic vector
E；
E=[E1,E2,...,Em-1,Em]
Wherein, EiThe energy value of node is represented, x represents the wavelet coefficient values in node, and wherein i represents nodes, and j is represented
Coefficient number in node；
34) effective node is analyzed, using characteristic vector Y obtain each node local entropy be distributed, obtain feature to
Measure F；
fi=pi·log(pi)
F=[f1,f2,...,fm-1,fm]
Wherein, piRepresent the probability of node, fiRepresent node entropy, the vector that F is made up of each entropy；
35) feature vector, X is merged, E, F obtain Y, characteristic vector Y be mapped to higher-dimension by kernel function, according to correlation
Yojan and conversion are carried out to Y；
Y=[Xnew,E,F]
Ynew=[α '1,...,α'm][k(Y1,Ynew),...,k(Ym,Ynew)]T
Wherein, YnewThe Y-direction amount after dimensionality reduction is represented, α represents the characteristic vector of nuclear matrix；
36) Y is analyzed using prior distribution, obtains matrix and matrix in class between Y class, pass through separable measures value
Carry out the separability of evaluating characteristic vector；
JF=tr (Sw -1Sb)
Wherein, JFRepresent separable measures value, SwRepresent Scatter Matrix in class, SbRepresent class scatter matrix.
Preferably, being specifically included according to correlation progress yojan with conversion in the step 3：
A. nuclear equation K (x, y) calculating matrix K is utilized；
B. unit character vector corresponding to preceding multiple eigenvalue of maximum of matrix K is sought；
C. characteristic vector corresponding to choosing multiple eigenvalue of maximum brings equations projection coordinate into, using gaussian kernel function
It is as follows：
Wherein, x, xiRepresent different samples, σ representative sample mean square errors.
Preferably, quantitative classification identification step specifically includes in the step 4：
41) optimization of parameters is carried out to SVM classifier using based on the optimizing of scope, and training is obtained accurately with cross validation
Rate；
42) index of the supporting vector accounting as model generalization ability is used；
43) predict to obtain the average accuracy of sample using cross validation, average recall rate and F1 values；
TP --- positive class is predicted into positive class；
FN --- positive class is predicted into negative class；
FP --- negative class is predicted into positive class；
TN --- negative class is predicted into negative class；
Accurate rate：
Recall rate：
F1 values：
44) TFR of sample predictions, the performance of FPR classification of assessment are obtained using cross validation.
Beneficial effects of the present invention：
The present invention can reduce those of ordinary skill's mistake in different in width ultrasonic Flaw is identified, by being computer depth
Degree excavates signal in the feature and information of time-frequency domain, aided detection personnel.The program can realize that the artificial of 0.5mm width differences lacks
Fall into identification.
The method of the present invention need not tangle the ultrasonic reflection modeling in complexity, have preferable versatility and practicality.
The present invention is for realizing that online automatic ultrasonic detection has with evaluation, the efficiency of raising Ultrasonic NDT and quality
There is important meaning.
Brief description of the drawings
Fig. 1 is the structural representation of laser-ultrasound experimental system.
Fig. 2 is the profile and scale diagrams of test specimen.
Fig. 3 a are the performance schematic diagram of experiment primary signal.
Fig. 3 b are the performance schematic diagram of biol.5 small echos.
Fig. 3 c are the performance schematic diagram of sym3 small echos.
Fig. 3 d are the performance schematic diagram of coif3 small echos.
Fig. 3 e are the performance schematic diagram of demy small echos.
Fig. 4 a are the wide defect wavelet energy figures of 0.5mm.
Fig. 4 b are the wide defect wavelet energy figures of 1mm.
Fig. 5 is two-dimensional projection's schematic diagram of sample point.
Fig. 6 is tripleplane's schematic diagram of sample point.
Embodiment
For the ease of the understanding of those skilled in the art, the present invention is made further with reference to embodiment and accompanying drawing
Bright, the content that embodiment refers to not is limitation of the invention.
A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm of the present invention, with reference to shown in Fig. 1
Laser-ultrasound experimental system the test specimen shown in Fig. 2 is tested.
Laser-ultrasound experimental system includes：Nd:YAG laser, scanning galvanometer, controller, laser interferometer and data are adopted
Collecting system, data collecting system include amplifier and AD capture cards, PC；On PC by controller come control scanning galvanometer with
Laser sets the direction of illumination of laser beam and position, and caused ultrasonic wave is received by laser interferometer in surface of test piece, lead to
Amplifier and AD capture cards are crossed, signal is stored on PC.
Collection experiment echo-signal, processing step are as follows：
11) echo position is calculated by velocity of wave according to defect and the distance of shot point；
12) echo waveform is intercepted.
The signal collected is subjected to denoising, using five layers of wavelet soft-threshold denoising, its threshold value uses rigrsure
Threshold value, the threshold values are to use the adaptive threshold selection based on Stein unbiaseds/possibility predication principle, give a threshold values λ, obtain
Minimized to its possibility predication, then by non-likelihood λ, so that it may obtain selected threshold values.Signal is gone using four kinds of orthogonal wavelets
Make an uproar, most suitable coif3 small echos are chosen according to signal to noise ratio and mean square error, as shown in Fig. 3 a- Fig. 3 e.
Five layers of demy wavelet package transforms are carried out to signal, and choose the effective node of signal, as shown in Fig. 4 a, Fig. 4 b.
Using as follows to the signal transacting after packet transform based on the characteristics algorithm of wavelet packet fusion：
31) sub-band coefficients are obtained using wavelet package transforms and calculates the Energy distribution of each node, choose effective node；
X=[x00,x01,x02,...,xij,xij+1,...,xmn]
Vectorial X includes all wavelet packet coefficients, wherein, i represents nodes, and j represents the coefficient number in node, m i
Sum, n be j sum；
32) correlation analysis is carried out to effective node：Data are mapped to higher dimensional space by kernel function, further according to correlation
Property yojan and conversion carried out to this class sample obtain feature vector, X；
Xnew=[α1,...,αm][k(X1,Xnew),...,k(Xm,Xnew)]T
Wherein, α is the characteristic vector of nuclear matrix, and k is nuclear matrix；
33) effective node is analyzed, obtains the energy Spectral structure of each node using frequency spectrum algorithm, obtain characteristic vector
E；
E=[E1,E2,...,Em-1,Em]
Wherein, EiThe energy value of node is represented, x represents the wavelet coefficient values in node, and i represents nodes, and j represents node
In coefficient number；
34) effective node is analyzed, using characteristic vector Y obtain each node local entropy be distributed, obtain feature to
Measure F；
fi=pi·log(pi)
F=[f1,f2,...,fm-1,fm]
Wherein, piRepresent the probability of node, fiRepresent node entropy, the vector that F is made up of each entropy；
35) feature vector, X is merged, E, F obtain Y, characteristic vector Y be mapped to higher-dimension by kernel function, according to correlation
Yojan and conversion are carried out to Y；
Y=[Xnew,E,F]
Ynew=[α '1,...,α'm][k(Y1,Ynew),...,k(Ym,Ynew)]T
Wherein, YnewThe Y-direction amount after dimensionality reduction is represented, α represents the characteristic vector of nuclear matrix；
36) Y is analyzed using prior distribution, obtains matrix and matrix in class between Y class, pass through separable measures value
Carry out the separability of evaluating characteristic vector；
JF=tr (Sw -1Sb)
Wherein, JFRepresent separable measures value, SwRepresent Scatter Matrix in class, SbRepresent class scatter matrix.
Described specifically includes according to correlation progress yojan with conversion：
A. nuclear equation K (x, y) calculating matrix K is utilized；
B. unit character vector corresponding to preceding multiple eigenvalue of maximum of matrix K is sought；
C. characteristic vector corresponding to choosing multiple eigenvalue of maximum brings equations projection coordinate into, using gaussian kernel function
It is as follows：
Wherein, x, xiRepresent different samples, σ representative sample mean square errors.
In order to verify the validity of features described above extraction algorithm, sample point is projected into two-dimentional (such as Fig. 5) and three-dimensional (such as figure
6)。
The step of quantitative classification identifies is as follows：
41) optimization of parameters is carried out to SVM classifier using based on the optimizing of scope, and training is obtained accurately with cross validation
Rate；
42) index of the supporting vector accounting as model generalization ability is used；
43) predict to obtain the average accuracy of sample using cross validation, average recall rate and F1 values；
44) TFR of sample predictions, the performance of FPR classification of assessment are obtained using cross validation.
The method of the present invention, combines Non-Destructive Testing, machine learning and area of pattern recognition, from Fig. 4 a, Fig. 4 b, Fig. 5 institutes
Show, the present invention can preferably realize the quantitative detection of sample with reference to ultrasonic technique.
Concrete application approach of the present invention is a lot, and described above is only the preferred embodiment of the present invention, it is noted that for
For those skilled in the art, under the premise without departing from the principles of the invention, some improvement can also be made, this
A little improve also should be regarded as protection scope of the present invention.
Claims (6)
1. a kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm, it is characterised in that including step such as
Under：
Step 1：The surface wave signal of flaw echo is gathered, and carries out waveform interception；
Step 2：Gone using coif3 small echos dry；
Step 3：Defect characteristic is extracted using wavelet packet fusion feature extraction algorithm；
Step 4：Classification quantitative judge is carried out using sorting algorithm.
2. the ultrasonic quantitative detection method according to claim 1 based on wavelet packet fusion feature extraction algorithm, its feature
It is, the step 2 specifically includes：Using five layers of wavelet soft-threshold denoising.
3. the ultrasonic quantitative detection method according to claim 2 based on wavelet packet fusion feature extraction algorithm, its feature
It is, the step 2 specifically also includes：Five layers of wavelet soft-threshold use rigrsure threshold values, and the threshold values is that use is based on
The adaptive threshold selection of Stein unbiaseds/possibility predication principle, a threshold values λ is given, obtains its possibility predication, then will be non-
Likelihood λ is minimized, and obtains selected threshold values.
4. the ultrasonic quantitative detection method according to claim 1 based on wavelet packet fusion feature extraction algorithm, its feature
It is, the feature extraction algorithm in the step 3 specifically includes：
31) sub-band coefficients are obtained using wavelet package transforms and calculates the Energy distribution of each node, choose effective node；
32) correlation analysis is carried out to effective node, data is mapped to higher dimensional space by kernel function, further according to correlation pair
This class sample carries out yojan and obtains feature vector, X with conversion；
33) effective node is analyzed, obtains the energy Spectral structure of each node using frequency spectrum algorithm, obtain characteristic vector E；
34) effective node is analyzed, the local entropy that each node is obtained using characteristic vector Y is distributed, and obtains characteristic vector F；
35) feature vector, X is merged, E, F obtain Y, characteristic vector Y are mapped to higher-dimension by kernel function, Y entered according to correlation
Row yojan and conversion；
36) Y is analyzed using prior distribution, obtains matrix and matrix in class between Y class, commented by separable measures value
The separability of valency characteristic vector.
5. the ultrasonic quantitative detection method according to claim 4 based on wavelet packet fusion feature extraction algorithm, its feature
It is, being specifically included according to correlation progress yojan with conversion in the step 3：
A. nuclear equation K (x, y) calculating matrix K is utilized；
B. unit character vector corresponding to preceding multiple eigenvalue of maximum of matrix K is sought；
C. characteristic vector corresponding to choosing multiple eigenvalue of maximum brings equations projection coordinate into, using gaussian kernel function such as
Under：
<mrow>
<mi>K</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mfrac>
<mrow>
<mo>|</mo>
<mo>|</mo>
<mi>x</mi>
<mo>-</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>|</mo>
<mo>|</mo>
</mrow>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
Wherein, x, xiRepresent different samples, σ representative sample mean square errors.
6. the ultrasonic quantitative detection method according to claim 1 based on wavelet packet fusion feature extraction algorithm, its feature
It is, quantitative classification identification step specifically includes in the step 4：
41) optimization of parameters is carried out to SVM classifier using based on the optimizing of scope, and training accuracy rate is obtained with cross validation；
42) index of the supporting vector accounting as model generalization ability is used；
43) predict to obtain the average accuracy of sample using cross validation, average recall rate and F1 values；
44) TFR of sample predictions, the performance of FPR classification of assessment are obtained using cross validation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710760358.8A CN107727749B (en) | 2017-08-30 | 2017-08-30 | Ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710760358.8A CN107727749B (en) | 2017-08-30 | 2017-08-30 | Ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107727749A true CN107727749A (en) | 2018-02-23 |
| CN107727749B CN107727749B (en) | 2020-08-07 |
Family
ID=61204791
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710760358.8A Active CN107727749B (en) | 2017-08-30 | 2017-08-30 | Ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107727749B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109975136A (en) * | 2019-04-09 | 2019-07-05 | 西南交通大学 | A kind of steel-frame structure damnification recognition method based on wavelet packet analysis |
| CN113984909A (en) * | 2021-11-03 | 2022-01-28 | 陕西省地方电力(集团)有限公司西安供电分公司 | Machine learning-based insulating equipment defect identification and imaging method |
| CN115034261A (en) * | 2022-05-26 | 2022-09-09 | 云南财经大学 | Method and equipment for extracting features between pulses of radar radiation source signal and storage medium |
| CN115294376A (en) * | 2022-04-24 | 2022-11-04 | 西京学院 | Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002345821A (en) * | 2001-05-29 | 2002-12-03 | Koji Hattori | Ultrasonic analytic system built in articular cavity |
| KR100497794B1 (en) * | 2002-07-26 | 2005-06-28 | 김희찬 | Ultrasonic inspection method and apparatus using wavelet ultrasonic wave |
| JP3837491B2 (en) * | 2001-02-02 | 2006-10-25 | 北陸電力株式会社 | Material damage detection method |
| JP4126817B2 (en) * | 1999-08-26 | 2008-07-30 | 株式会社Ｉｈｉ | Film thickness measuring method and apparatus |
| CN101620203A (en) * | 2009-05-13 | 2010-01-06 | 中华人民共和国南通出入境检验检疫局 | Deficiency ultrasonic detecting equipment of mechanical equipment based on wavelet theory |
| JP4591850B2 (en) * | 2005-01-18 | 2010-12-01 | 株式会社Ｉｈｉ検査計測 | Ultrasonic inspection method and apparatus |
| JP4606860B2 (en) * | 2004-12-03 | 2011-01-05 | 川崎重工業株式会社 | Defect identification method and apparatus by ultrasonic inspection |
| CN102082754A (en) * | 2009-11-26 | 2011-06-01 | 中兴通讯股份有限公司 | OFDM (Orthogonal Frequency Division Multiplexing) channel estimation method and device |
| JPWO2011013802A1 (en) * | 2009-07-31 | 2013-01-10 | 国立大学法人九州大学 | Nondestructive inspection method and apparatus |
| JP5507279B2 (en) * | 2010-02-02 | 2014-05-28 | 株式会社Ｉｈｉ検査計測 | Ultrasonic inspection method and apparatus |
| US20140236500A1 (en) * | 2011-09-29 | 2014-08-21 | Ge Sensing & Inspection Technologies Gmbh | Method for processing an ultrasonic analog signal, digital signal processing unit and ultrasonic inspection device |
| CN104714237A (en) * | 2015-01-30 | 2015-06-17 | 哈尔滨工程大学 | Fish identification method with multi-feature and multidirectional data fused |
| CN105137498A (en) * | 2015-09-17 | 2015-12-09 | 鲁东大学 | Underground target detection and recognition system and method based on feature fusion |
| CN106017879A (en) * | 2016-05-18 | 2016-10-12 | 河北工业大学 | Universal circuit breaker mechanical fault diagnosis method based on feature fusion of vibration and sound signals |
| CN106093120A (en) * | 2016-05-27 | 2016-11-09 | 北京科技大学 | A kind of Feature Parameter Fusion method of Electronic Nose response signal in tea grades identification |
| CN106909926A (en) * | 2015-12-21 | 2017-06-30 | 深圳市科彩印务有限公司 | A kind of uniformity detecting method and device of cigarette bag printing color |
-
2017
- 2017-08-30 CN CN201710760358.8A patent/CN107727749B/en active Active
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4126817B2 (en) * | 1999-08-26 | 2008-07-30 | 株式会社Ｉｈｉ | Film thickness measuring method and apparatus |
| JP3837491B2 (en) * | 2001-02-02 | 2006-10-25 | 北陸電力株式会社 | Material damage detection method |
| JP2002345821A (en) * | 2001-05-29 | 2002-12-03 | Koji Hattori | Ultrasonic analytic system built in articular cavity |
| KR100497794B1 (en) * | 2002-07-26 | 2005-06-28 | 김희찬 | Ultrasonic inspection method and apparatus using wavelet ultrasonic wave |
| JP4606860B2 (en) * | 2004-12-03 | 2011-01-05 | 川崎重工業株式会社 | Defect identification method and apparatus by ultrasonic inspection |
| JP4591850B2 (en) * | 2005-01-18 | 2010-12-01 | 株式会社Ｉｈｉ検査計測 | Ultrasonic inspection method and apparatus |
| CN101620203A (en) * | 2009-05-13 | 2010-01-06 | 中华人民共和国南通出入境检验检疫局 | Deficiency ultrasonic detecting equipment of mechanical equipment based on wavelet theory |
| JPWO2011013802A1 (en) * | 2009-07-31 | 2013-01-10 | 国立大学法人九州大学 | Nondestructive inspection method and apparatus |
| CN102082754A (en) * | 2009-11-26 | 2011-06-01 | 中兴通讯股份有限公司 | OFDM (Orthogonal Frequency Division Multiplexing) channel estimation method and device |
| JP5507279B2 (en) * | 2010-02-02 | 2014-05-28 | 株式会社Ｉｈｉ検査計測 | Ultrasonic inspection method and apparatus |
| US20140236500A1 (en) * | 2011-09-29 | 2014-08-21 | Ge Sensing & Inspection Technologies Gmbh | Method for processing an ultrasonic analog signal, digital signal processing unit and ultrasonic inspection device |
| CN104714237A (en) * | 2015-01-30 | 2015-06-17 | 哈尔滨工程大学 | Fish identification method with multi-feature and multidirectional data fused |
| CN105137498A (en) * | 2015-09-17 | 2015-12-09 | 鲁东大学 | Underground target detection and recognition system and method based on feature fusion |
| CN106909926A (en) * | 2015-12-21 | 2017-06-30 | 深圳市科彩印务有限公司 | A kind of uniformity detecting method and device of cigarette bag printing color |
| CN106017879A (en) * | 2016-05-18 | 2016-10-12 | 河北工业大学 | Universal circuit breaker mechanical fault diagnosis method based on feature fusion of vibration and sound signals |
| CN106093120A (en) * | 2016-05-27 | 2016-11-09 | 北京科技大学 | A kind of Feature Parameter Fusion method of Electronic Nose response signal in tea grades identification |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109975136A (en) * | 2019-04-09 | 2019-07-05 | 西南交通大学 | A kind of steel-frame structure damnification recognition method based on wavelet packet analysis |
| CN109975136B (en) * | 2019-04-09 | 2021-06-25 | 西南交通大学 | Steel frame structure damage identification method based on wavelet packet analysis |
| CN113984909A (en) * | 2021-11-03 | 2022-01-28 | 陕西省地方电力(集团)有限公司西安供电分公司 | Machine learning-based insulating equipment defect identification and imaging method |
| CN115294376A (en) * | 2022-04-24 | 2022-11-04 | 西京学院 | Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics |
| CN115034261A (en) * | 2022-05-26 | 2022-09-09 | 云南财经大学 | Method and equipment for extracting features between pulses of radar radiation source signal and storage medium |
| CN115034261B (en) * | 2022-05-26 | 2023-08-22 | 云南财经大学 | Method, equipment and storage medium for extracting inter-pulse characteristics of radar radiation source signals |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107727749B (en) | 2020-08-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112858473B (en) | Turnout switch blade damage state monitoring method based on feature fusion | |
| CN112232400B (en) | Stainless steel weld ultrasonic defect detection method based on depth feature fusion | |
| Dorafshan et al. | Deep learning models for bridge deck evaluation using impact echo | |
| CN112462346A (en) | Ground penetrating radar roadbed defect target detection method based on convolutional neural network | |
| CN107727749A (en) | A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm | |
| CN104359389B (en) | Pulsed eddy current measuring method for measuring relative wall thickness variable of ferromagnetic component | |
| CN105334269A (en) | Pipeline defect type determination method based on neural network and guided wave characteristic database | |
| CN107121501A (en) | A kind of turbine rotor defect classification method | |
| CN104764413B (en) | marine structure deck plate welding deformation measuring method | |
| CN111539489A (en) | Millimeter wave image target detection and identification method | |
| Vy et al. | Damage localization using acoustic emission sensors via convolutional neural network and continuous wavelet transform | |
| Divós et al. | Resolution of stress wave based acoustic tomography | |
| CN103645248B (en) | A kind of high-temperature alloy grain size evaluation method based on ultrasonic phase velocity | |
| Merazi Meksen et al. | Automatic crack detection and characterization during ultrasonic inspection | |
| CN110208377A (en) | A kind of more characteristic parameters damage degree assessment method based on Lamb wave | |
| CN107356677B (en) | Ultrasonic nondestructive testing method based on travel time tomography and reverse time migration imaging | |
| CN1743839A (en) | Structure defect ultrasonic on-line intelligent identifying system and identifying method | |
| CN112131781B (en) | Steel structure damage detection method based on fully-connected neural network and transfer rate function | |
| CN104749249B (en) | A kind of method of the detection physical purity of seed based on ultrasonic technology | |
| CN113468804B (en) | Underground pipeline identification method based on matrix bundles and deep neural network | |
| Li et al. | Non-destructive detection on the egg crack based on wavelet transform | |
| CN117233347B (en) | Carbon steel spheroidization grade measuring method, system and equipment | |
| CN112836274A (en) | Data fusion method for tracing and auditing hidden engineering | |
| CN105891339B (en) | Utilize the method, apparatus and system of impact Imaging Method detection geotechnical engineering media defect | |
| CN117349718A (en) | Anti-noise damage classification method for composite material plate structure based on Lamb waves |
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 |