CN107886056A - A kind of electronic nose of fuzzy covariance learning network differentiates vinegar kind method - Google Patents
A kind of electronic nose of fuzzy covariance learning network differentiates vinegar kind method Download PDFInfo
- Publication number
- CN107886056A CN107886056A CN201711027258.0A CN201711027258A CN107886056A CN 107886056 A CN107886056 A CN 107886056A CN 201711027258 A CN201711027258 A CN 201711027258A CN 107886056 A CN107886056 A CN 107886056A
- Authority
- CN
- China
- Prior art keywords
- vinegar
- class
- fuzzy
- iteration
- electronic nose
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Electrochemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Analytical Chemistry (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- General Engineering & Computer Science (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
Abstract
The invention discloses a kind of electronic nose of fuzzy covariance learning network to differentiate vinegar kind method, specifically includes following steps:First by the vinegar data of electronic nose collection different cultivars, then (SNV) is converted by standard normal variable to pre-process the vinegar data collected, data compression being carried out with principal component analysis (PCA) again and the authentication information of training sample being extracted in linear discriminant analysis (LDA), finally vinegar kind is differentiated according to a kind of method of fuzzy covariance learning network.The present invention has classification accuracy high, and easy, lossless, cost is low, easy to implement and popularization and application.
Description
Technical field
The present invention relates to vinegar Variety identification field, and in particular to a kind of electronic nose of fuzzy covariance learning network differentiates
Vinegar kind method.
Background technology
Vinegar is one of flavouring of many families, there is various vinegars, the vinegar of different brands currently on the market
And different brewing methods and different brewing materials, then the quality of vinegar and flavor are not quite similar.As consumers in general are to food
The quality of vinegar has more requirements, and the differentiation research for vinegar quality and kind has important Research Significance and research valency
Value.
In general, the method to odor discrimination is to pass through training by some, and the people for having years of work experience comes
Carry out, and the objective determination method to smell is usually to use gas chromatography.But both approaches have shortcoming:Professional people
Member's subjectivity is higher, and culture is specialized in the professional of discriminating and cost dearly;Tested using gas chromatography, not only
Process is complicated, and his environmental requirement to experiment is very harsh, and carrying out Site Detection with gas chromatography has certain limitation
Property.Compared with both approaches, electronic nose has the advantages of notable, so being carried out using Electronic Nose Technology to vinegar quick, accurate
True qualitative analysis, there is very big application potential.
Zhang Houbo etc. obtains data from sensor matrices, and (Zhang Houbo, Mei Xiaodong, Zhao Wan, Wang Biao, Lu Ge spaces are used to eat
Electronic nose research [J] the sensors and micro-system of vinegar quality pre assessment, 2013,32 (3):62-64.), then to these data make
The electronic nose data of five kinds of vinegars are analyzed and handled with principal component analysis (PCA) and linear discriminant analysis (LDA), can be with
Preferably the vinegar of different cultivars is classified, a kind of easily method is provided for the pre assessment of vinegar quality.PCA is main
It is that the electronic nose data of multidimensional are subjected to dimensionality reduction, but dimensionality reduction is the compression of data, lifting vinegar that can not be apparent divides
The accuracy rate of class.LDA is used for the discriminant analysis information for extracting electronic nose data, and the accuracy rate classified for lifting vinegar has very big
Help;It is not a kind of grader but LDA is also a kind of Data Dimensionality Reduction processing method.
2014, the electric nasus system that Wang Weiwei et al. devises PCA and BP hybrid neural networks algorithms (Wang Weiwei, was opened
Match electric nasus system [J] sensor and micro-system of the male based on PCA Yu BP hybrid neural networks algorithms, 2014,33 (4):90-
94.), first with PCA by the electronic nose Data Dimensionality Reduction of vinegar, then the new variables obtained by PCA is carried out with BP neural network
Pattern-recognition is so as to realizing the classification of vinegar kind.Although neutral net (BP) be able to can be realized non-thread by learning sample data
Property classification, but neutral net exist local minimum point, cross study the problems such as.
The content of the invention
The problem of existing for electronic nose sorting technique used in current vinegar assortment, the present invention propose a kind of fuzzy
The electronic nose of covariance learning network differentiates vinegar kind method, this method provide one kind simply, easily and fast, accurately
Electronic nose vinegar assortment method.
The principle of foundation of the present invention:The gas that tested vinegar sample volatilizes acts on electronic nose sensor array, causes
The change of sensor output voltage;The vinegar sample of different cultivars, because its vinegar brewing method and the phase not to the utmost of brewing materials
Together, so as to cause its smell volatilized difference to be present, thus electronic nose sensor output voltage is caused to have differences, according to this
The difference classification that suitable data processing method is the vinegar sample that different cultivars can be achieved.
A kind of electronic nose vinegar assortment method of fuzzy covariance learning network, specifically includes following steps:
S1, the vinegar sample of different cultivars is gathered using electric nasus system;
S2, pre-processed with standard normal variable conversion process, then with principal component analysis (PCA) carry out data compression and
The authentication information of training sample is extracted in linear discriminant analysis (LDA);PCA carries out 6 dimension datas that data compression obtains and is divided into training
Sample and test sample, training sample are used for pattern drill, and test sample is used for test mode recognition correct rate;
S3, vinegar assortment is carried out with a kind of fuzzy covariance learning network, is comprised the following steps that:
S3.1, parameter initialization:It is c (+∞ > c >=2) to set vinegar kind number, initial weight exponent m0(+∞ > m0>
1), greatest iteration number rmax, error higher limit ε, number of training n1, test sample number n2, initial classes center and initial is set
Fuzzy membership;
S3.2, calculate r-1 (r=1,2 ..., rmax) secondary iteration when apart from norm
Wherein For the r-1 times repeatedly
For when kth (k=1,2,3 ..., n2) individual test sample xkTo class center vi,r-1Distance, Ai,rI-th when being the r times iteration
The norm matrix of cluster centre, d be test sample dimension, vi,r-1For the r-1 times iteration when the i-th class class center (i=1,
2 ... ..., c), uik,r-1For the r-1 times iteration when test sample xkBelong to the fuzzy membership of the i-th class, Sfi,rIt is the r times iteration
When the i-th class fuzzy covariance matrix;
S3.3, calculate fuzzy membership angle value u during the r times iterationik,r;
Wherein it is subordinate to angle value uik,rRepresent k-th of sample during the r times iterative calculation
Originally it is under the jurisdiction of the angle value that is subordinate to of the i-th class, mrFor the r times iteration when weighted index, and mr=m0- r Δ m, Δ m=(m0-1)/
rmax;
S3.4, calculate learning rate α during the r times iterationik,r,
S3.5, calculate class center v during the r times iterationi,r
Wherein vi,rThe class center of i-th class, v when being iterated to calculate for the r timesi,r-1For
The class center of i-th class during the r-1 times iterative calculation;
S3.6, work as maxi||vi,r-vi,r-1| | < ε or r=rmaxWhen -1, iteration terminates, and otherwise returns to S3.2, continues to change
In generation, calculates;After iteration convergence, according to final fuzzy membership uik,rDifferentiate test sample xkWhich kind of belongs to, i.e. xkBelong to
The vinegar of which kind.
The beneficial effects of the invention are as follows:
1st, the present invention carries out the classification processing of vinegar e-nose signal using fuzzy covariance learning network, due to using mould
Paste degree of membership is classified so that is better than conventional sorting methods in terms of Noise e-nose signal is handled.
2nd, the present invention using fuzzy covariance matrix apart from norm, for the e-nose signal of complicated classification distribution shape
With advantage, classification accuracy is greatly improved.
Brief description of the drawings
Fig. 1 is the flow chart of the inventive method;
Fig. 2 is the datagram that vinegar sample obtains after standard normal variable conversion process (SNV) processing;
Fig. 3 is the 6 dimension data figures that vinegar sample obtains after PCA compresses;
Fig. 4 is the datagram that test sample obtains after LDA is handled;
Fig. 5 is initial fuzzy membership figure;
Fig. 6 is the fuzzy membership figure after iteration convergence.
Embodiment
The device and method of the present invention is described further with reference to the accompanying drawings and detailed description.
As shown in figure 1, the method key step for vinegar assortment includes:
Step 1: the vinegar sample of different cultivars is gathered using electric nasus system.
In the environment of 20 DEG C of room temperature, humidity 40%, electronic nose is powered, after sensor preheats 10 minutes, into beaker
Pour into 10ml vinegars and put it into casing, cover case lid, timing rapidly;Respectively 60 minutes, 65 minutes, 70 minutes three
The host computer procedure that individual time point is write using labview carries out electronic nose data acquisition, takes the average value of collection result three times
As a vinegar sample data;After completing a vinegar sample collection, open case lid and each sensor restPosed,
Then repeated acquisition vinegar sample processes.
Data explanation:To 5 class vinegars (Zhenjiang mature vinegar, zhenjiang vinegar, Shanxi mature vinegar, fixed speed wind turbine, protecting peaceful vinegar) in this example,
There are 51 samples per class vinegar, collect 255 vinegar electronic nose sample datas altogether, experiment gathered data result is preserved;
Each sample is the vector of 1 × 10 (once experiment obtains ten sensor response datas), and total sample is 255 × 10 data square
Battle array.
Step 2: being pre-processed with standard normal variable conversion process (SNV), then carried out with principal component analysis (PCA)
Data compression and the authentication information of linear discriminant analysis (LDA) extraction training sample.
1st, the initial data of step 1 is pre-processed and carries out data compression with principal component analysis (PCA)
Raw data matrix (255x10) obtained by from step 1 is carried out standard normalized, is as a result such as figure
Shown in 2;Then handled by carrying out principal component analysis (PCA), 6 characteristic values before PCA are calculated:39.6945、28.7220、
10.5444th, 4.0413, the 0.4454, data matrix that 0.2454, PCA characteristic vector is 10 × 6, as shown in table 1;Total sample
6 dimension datas as shown in Figure 3 are obtained after PCA compresses for 255 × 10 data matrix, that is, are transformed to 255 × 6 data square
Battle array.
The PCA of table 1 characteristic vector
1 | 2 | 3 | 4 | 5 | 6 | |
1 | -0.1937 | -0.7929 | 0.0310 | 0.0535 | 0.1594 | 0.0566 |
2 | -0.2443 | -0.1355 | -0.3081 | -0.3660 | 0.1772 | 0.4185 |
3 | -0.0382 | 0.1644 | -0.6384 | -0.28057 | -0.46862 | -0.24338 |
4 | -0.17566 | 0.018888 | 0.22744 | 0.091881 | -0.14014 | -0.22585 |
5 | 0.18328 | 0.14437 | -0.397 | 0.78733 | 0.20363 | 0.15509 |
6 | -0.34556 | 0.54267 | 0.26536 | -0.12548 | 0.26147 | 0.36597 |
7 | -0.16354 | 0.027358 | 0.2485 | 0.11369 | -0.21355 | -0.26827 |
8 | -0.16887 | 0.031488 | 0.25584 | 0.13074 | -0.11879 | -0.30122 |
9 | 0.62227 | -0.066153 | 0.29888 | -0.083701 | -0.44253 | 0.4627 |
10 | 0.5243 | 0.0664 | 0.0164 | -0.3214 | 0.5818 | -0.4202 |
6 dimension datas after PCA is handled are divided into training sample and test sample, by 25 in 51 samples of every class vinegar
As training sample, remaining 26 samples are as test sample;Training sample is used for pattern drill, test sample is used for examining
Test pattern-recognition accuracy;Training sample sum is n1=125, each sample is 1 × 6 vector, obtains 125 × 6 number
According to matrix;Test sample sum is n2=130, each sample is 1 × 6 vector, obtains 130 × 6 data matrix.
2nd, the authentication information of training sample is extracted with linear discriminant analysis (LDA)
By using the authentication information of the linear discriminant analysis extraction data matrix of training sample 125 × 6, it is calculated
4 characteristic values are before LDA:118.0375th, 14.0824,9.4818,0.6353, the number of 4 LDA discriminant vectorses composition 6 × 4
According to matrix, as shown in table 2;The data matrix of test sample 130 × 6 is projected on LDA discriminant vectorses after can obtaining conversion
Test sample be as Fig. 4 130 × 4 data matrix.
The LDA of table 24 discriminant vectorses
1 | 2 | 3 | 4 | |
1 | -0.2382 | -0.2553 | -0.0149 | -0.0052 |
2 | -0.7636 | 0.1297 | -0.0228 | -0.0153 |
3 | -0.1147 | 0.0765 | -0.3359 | 0.2219 |
4 | 0.4322 | -0.0776 | -0.5269 | -0.1486 |
5 | -0.3631 | -0.8472 | 0.5858 | 0.8809 |
6 | -0.1677 | 0.4338 | 0.5152 | 0.3902 |
Step 3: carrying out vinegar assortment with a kind of fuzzy covariance learning network, comprise the following steps that:
A, initialize:It is c (+∞ > c >=2) to set vinegar kind number, initial weight exponent m0(+∞ > m0> 1), it is maximum
Number of iterations rmax, the value ε of the error upper limit, number of training n1, test sample number n2, initial classes center v is seti,0And introductory die
Paste degree of membership uik,0;
Design parameter is arranged to:Vinegar kind number is c=5, initial weight exponent m0=2, greatest iteration number rmax=100,
Value ε=0.00001 of the error upper limit, number of training n1=125, test sample number n2=130.
Initial classes center vi,0With initial fuzzy membership uil,0It is calculated as follows:
vi,0=x_mean (i) (1)
Using training sample per class average as the initial classes center per class, x_mean (i) is i-th (i=1,2,3,4,5) class
Training sample average, vi,0For the initial classes center of the i-th class, vj,0For the initial classes center of jth (j=1,2,3,4,5) class;
uil,0For l (l=1,2,3 ..., n1) individual training sample xlIt is under the jurisdiction of the initial fuzzy membership of the i-th class, uil,0Such as Fig. 5 institutes
Show.
B, calculate r-1 (r=1,2 ..., rmax) secondary iteration when apart from norm
Wherein:
In above formulaFor the r-1 times iteration when kth (k=1,2,3 ..., n2) individual test sample xkTo class center vi,r-1
Distance, Ai,rThe norm matrix at ith cluster center when being the r times iteration;D is the dimension of test sample;xkFor k-th
Test sample, vi,r-1For the r-1 times iteration when the i-th class class center (i=1,2 ... ..., c), uik,r-1For the r-1 times iteration when
Test sample xkBelong to the fuzzy membership of the i-th class, Sfi,rThe fuzzy covariance matrix of i-th class when being the r times iteration.
C, r (r=1,2 ... ..., r are calculatedmax) secondary iteration when fuzzy membership angle value uik,r;
It is subordinate to angle value uik,rRepresent r (r=1,2 ... ..., rmax) secondary iterative calculation when k-th of test sample be under the jurisdiction of
I classes are subordinate to angle value, mrFor the r times iteration when weighted index, mr=m0-rΔm;Δ m=(m0-1)/rmax;
D, learning rate α during the r times iteration is calculatedik,r
E, class center v during the r times iteration is calculatedi,r(i=1,2 ... ..., c)
Wherein vi,rThe class center of i-th (i=1,2 ... ..., c) class, v when being iterated to calculate for the r timesi,r-1For the r-1 times repeatedly
The class center of i-th class when generation calculates;
F, max is worked asi||vI, r-vI, r-1| | < ε or r=rmaxWhen -1, iteration terminates, and otherwise return to step B continues iteration
Calculate.After iteration convergence, test sample x is differentiated according to final fuzzy membershipkWhich kind of belongs to, i.e. xkWhich product belonged to
The vinegar of kind.
Result of calculation:Fuzzy membership u after iteration convergenceik,rIf as shown in fig. 6, uik,rThe then discriminating test samples of > 0.5
xkBelong to the i-th class;The classification accuracy of vinegar kind test sample can be obtained up to 100% according to Fig. 6 fuzzy membership.
Although an embodiment of the present invention has been shown and described, it will be understood by those skilled in the art that:Not
In the case of departing from the principle and objective of the present invention a variety of change, modification, replacement and modification can be carried out to these embodiments, this
The scope of invention is limited by claim and its equivalent.
Claims (8)
1. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network, it is characterised in that comprise the following steps:
S1, the vinegar sample of different cultivars is gathered using electric nasus system;
S2, pre-processed with standard normal variable conversion process, then carry out data compression with principal component analysis PCA and linearly sentence
Not Fen Xi LDA extraction training sample authentication information;
S3, vinegar assortment is carried out with fuzzy covariance learning network.
2. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network as claimed in claim 1, its feature
It is, the S3 is specially:
S3.1, parameter initialization;
S3.2, calculate the r-1 times iteration when apart from normR=1,2 ..., rmax, rmaxFor greatest iteration number;
S3.3, calculate fuzzy membership angle value u during the r times iterationik,r;
S3.4, calculate learning rate α during the r times iterationik,r;
S3.5, calculate class center v during the r times iterationi,r, i=1,2 ... ..., c, c is vinegar kind number;
S3.6, work as maxi||vi,r-vi,r-1| | < ε or r=rmaxWhen -1, iteration terminates, and otherwise returns to S3.2, continues iteration meter
Calculate;After iteration convergence, test sample x is differentiated according to final fuzzy membershipkWhich kind of belongs to, i.e. xkWhich kind belonged to
Vinegar;Wherein vi,rThe class center of i-th class, v when being iterated to calculate for the r timesi,r-1When being iterated to calculate for the r-1 times in the class of the i-th class
The heart, ε are error higher limit.
3. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network as claimed in claim 2, its feature
It is, the S3.1 is specially:It is c to set vinegar kind number, initial weight exponent m0, greatest iteration number rmax, error higher limit
ε, number of training n1, test sample number n2, initial classes center v is seti,0With initial fuzzy membership uik,0;Wherein+∞ > c
>=2 ,+∞ > m0> 1.
4. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network as claimed in claim 2 or claim 3, it is special
Sign is, during the r-1 times iteration apart from normFor:Wherein To be surveyed for k-th during the r-1 times iteration
This x of samplekTo class center vi,r-1Distance, k=1,2,3 ..., n2, Ai,rThe norm at ith cluster center when being the r times iteration
Matrix, d be test sample dimension, vi,r-1For the r-1 times iteration when the i-th class class center, i=1,2 ... ..., c, uik,r-1For
K-th of test sample x during the r-1 times iterationkBelong to the fuzzy membership of the i-th class, Sfi,rThe i-th class is fuzzy when being the r times iteration
Covariance matrix.
5. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network as claimed in claim 2 or claim 3, it is special
Sign is, the fuzzy membership angle value uik,rFor: Wherein it is subordinate to angle value uik,rRepresent
K-th of sample is under the jurisdiction of the angle value that is subordinate to of the i-th class, m during the r times iterative calculationrFor the r times iteration when weighted index, and mr=
m0- r Δ m, Δ m=(m0-1)/rmax。
6. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network as claimed in claim 2, its feature
It is, the learning rate αik,rFor:Wherein mrFor the r times iteration when weighted index.
7. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network as claimed in claim 2, its feature
It is, class center v during the r times iterationi,rFor:Wherein vi,rFor the r times iterative calculation
When the i-th class class center, vi,r-1The class center of i-th class when being iterated to calculate for the r-1 times.
8. a kind of electronic nose vinegar assortment method of fuzzy covariance learning network as claimed in claim 1, its feature
It is, 6 dimension datas that PCA progress data compression obtains in the S2 are divided into training sample and test sample, and training sample is used for
Pattern drill, test sample are used for test mode recognition correct rate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711027258.0A CN107886056A (en) | 2017-10-27 | 2017-10-27 | A kind of electronic nose of fuzzy covariance learning network differentiates vinegar kind method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711027258.0A CN107886056A (en) | 2017-10-27 | 2017-10-27 | A kind of electronic nose of fuzzy covariance learning network differentiates vinegar kind method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107886056A true CN107886056A (en) | 2018-04-06 |
Family
ID=61782618
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711027258.0A Pending CN107886056A (en) | 2017-10-27 | 2017-10-27 | A kind of electronic nose of fuzzy covariance learning network differentiates vinegar kind method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107886056A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378375A (en) * | 2019-06-12 | 2019-10-25 | 江苏大学 | A kind of vinegar classification method of fuzzy discrete degree cluster |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104077598A (en) * | 2014-06-27 | 2014-10-01 | 电子科技大学 | Emotion recognition method based on speech fuzzy clustering |
US20160086047A1 (en) * | 2014-09-24 | 2016-03-24 | Northeastern University | Apparatus and method for extracting feature of image including object |
CN106951914A (en) * | 2017-02-22 | 2017-07-14 | 江苏大学 | The Electronic Nose that a kind of Optimization of Fuzzy discriminant vectorses are extracted differentiates vinegar kind method |
CN107192686A (en) * | 2017-04-11 | 2017-09-22 | 江苏大学 | A kind of Possibility Fuzzy Clustering local tea variety discrimination method of fuzzy covariance matrix |
-
2017
- 2017-10-27 CN CN201711027258.0A patent/CN107886056A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104077598A (en) * | 2014-06-27 | 2014-10-01 | 电子科技大学 | Emotion recognition method based on speech fuzzy clustering |
US20160086047A1 (en) * | 2014-09-24 | 2016-03-24 | Northeastern University | Apparatus and method for extracting feature of image including object |
CN106951914A (en) * | 2017-02-22 | 2017-07-14 | 江苏大学 | The Electronic Nose that a kind of Optimization of Fuzzy discriminant vectorses are extracted differentiates vinegar kind method |
CN107192686A (en) * | 2017-04-11 | 2017-09-22 | 江苏大学 | A kind of Possibility Fuzzy Clustering local tea variety discrimination method of fuzzy covariance matrix |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378375A (en) * | 2019-06-12 | 2019-10-25 | 江苏大学 | A kind of vinegar classification method of fuzzy discrete degree cluster |
CN110378375B (en) * | 2019-06-12 | 2023-08-15 | 江苏大学 | Vinegar classification method for fuzzy dispersion clustering |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2022160771A1 (en) | Method for classifying hyperspectral images on basis of adaptive multi-scale feature extraction model | |
CN108573225B (en) | Partial discharge signal pattern recognition method and system | |
CN106951914B (en) | Method for identifying vinegar variety by electronic nose for optimizing fuzzy identification vector extraction | |
CN105116397B (en) | Radar high resolution range profile target identification method based on MMFA models | |
CN106250848B (en) | A kind of single class Acoustic Object recognition methods based on multi-model fusion | |
CN109002859B (en) | Sensor array feature selection and array optimization method based on principal component analysis | |
CN109061426B (en) | Transformer partial discharge fault diagnosis method and online monitoring device | |
CN112149758B (en) | Hyperspectral open set classification method based on Euclidean distance and deep learning | |
CN114564982A (en) | Automatic identification method for radar signal modulation type | |
CN110378374B (en) | Tea near infrared spectrum classification method for extracting fuzzy identification information | |
CN104156628B (en) | A kind of naval vessel radiation signal recognition methods based on Multiple Kernel Learning discriminant analysis | |
CN109376805A (en) | A kind of classification method based on white wine base liquor Fingerprints | |
CN114091606A (en) | Tunnel blasting blast hole half-hole mark identification and damage flatness evaluation classification method | |
CN101067659B (en) | Remote sensing image sorting method | |
CN116204831A (en) | Road-to-ground analysis method based on neural network | |
CN108710836A (en) | A kind of lip detecting and read method based on cascade nature extraction | |
CN108764288A (en) | A kind of GK differentiates the local tea variety sorting technique of cluster | |
CN114596509A (en) | Machine vision-based rice seedling leaf age period identification method | |
Jain et al. | Mango leaf disease classification using deep learning hybrid model | |
CN107203779A (en) | Hyperspectral dimensionality reduction method based on spatial-spectral information maintenance | |
CN112183459B (en) | Remote sensing water quality image classification method based on evolution multi-objective optimization | |
CN107886056A (en) | A kind of electronic nose of fuzzy covariance learning network differentiates vinegar kind method | |
CN106570520A (en) | Infrared spectroscopy tea quality identification method mixed with GK clustering | |
CN106326914A (en) | SVM-based pearl multi-classification method | |
CN109409407A (en) | A kind of industry monitoring data clustering method based on LE algorithm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180406 |