CN104897608B - A kind of identification method for oolong quality based on near-infrared spectrum technique - Google Patents
A kind of identification method for oolong quality based on near-infrared spectrum technique Download PDFInfo
- Publication number
- CN104897608B CN104897608B CN201510343494.8A CN201510343494A CN104897608B CN 104897608 B CN104897608 B CN 104897608B CN 201510343494 A CN201510343494 A CN 201510343494A CN 104897608 B CN104897608 B CN 104897608B
- Authority
- CN
- China
- Prior art keywords
- spectrum
- tea
- oolong
- quality
- identification method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The present invention relates to a kind of identification method for oolong quality based on near-infrared spectrum technique, comprising the following steps: the preparation of tea powder sample;The acquisition of spectral information;The pretreatment of spectral information;Wavelength screening based on genetic algorithm;Identify the foundation of model and the measurement of quality of Oolong tea score.Sample pretreatment of the present invention is simple, can accurately and rapidly identify the quality of oolong tea, and this method strong operability avoids artificial sense and evaluates the influence of middle subjective consciousness and many and diverse operation of traditional chemical analysis method, as a result objective, accurate.
Description
Technical field
The present invention relates to a kind of identification method for oolong quality based on near-infrared spectrum technique.
Background technique
Oolong tea is one of big teas in China six, has distinct unique qualitative characteristics, is liked by the majority of consumers.With
The continuous expansion of oolong tea market scale, people have higher requirement to the quality of oolong tea, high-quality oolong tea is increasingly
It attracts attention.But the evaluation of current tea leaf quality relies primarily on traditional artificial sense and evaluates, as a result vulnerable to the person's of evaluating subjectivity because
The influence of element, and traditional chemical analysis method, cumbersome complexity, are not suitable for Tea Production pattern and sales of tea city
, this also limits the circulation and fair trade of China's tealeaves in the international market.
Therefore, the method that research tea leaf quality quickly detects seems especially urgent, to improve the inspection of China's tea leaf quality
Survey technology is horizontal, strengthens the differentiated control of tealeaves.
Summary of the invention
The purpose of the present invention is to provide a kind of identification method for oolong quality based on near-infrared spectrum technique, are oolong
Tea flavor evaluation provides a kind of fast and accurate approach, is conducive to standardization of the tea market realization to tea leaf quality management.
To achieve the above object, the present invention adopts the following technical scheme: a kind of oolong tea based on near-infrared spectrum technique
Quality identification method, it is characterised in that the following steps are included:
Step S1: the preparation of tea powder sample: the Tea Samples in each oolong tea place of production are collected, by the Tea Samples through crushing
Machine is ground into tea powder, and the tea powder is used as tea powder sample after sieving, number and encapsulation refrigeration;
Step S2: the acquisition of spectral information: the near infrared spectrum information collection work based on Workflow setting tea powder sample
It flows, acquires the tea powder sample in the spectrum information that diffuses near infrared range using near infrared spectrometer;
Step S3: the pretreatment of spectral information: collected to the step S2 unrestrained anti-using infrared processing software OPUS
It penetrates spectral information and carries out spectrum smoothing processing, second order derivation and normalized, while treated data are from spc spectrum format
Be converted to xls file format;
Step S4: the wavelength screening based on genetic algorithm: by all band spectrum of xls file format obtained in step S3
Data are equally divided into multiple subintervals and carry out random coded, and all various combinations in each subinterval constitute search space, use " R/
The maximum value of W " parameter filters out an optimized spectrum district's groups and closes, as the light spectral for participating in modeling as wavelength screening index;
Step S5: the foundation of model is identified: according to " GB/T23776-2009 tealeaves organoleptic evaluation method " to each tealeaves sample
Product carry out quality identification, and the processing for carrying out step S1 to step S4 to every a Tea Samples obtains corresponding light spectral,
Qualification result is established using Matlab mathematical software and handles obtained light spectral and identifies model correspondingly;
Step S6: the measurement of quality score: the spectrum that a Tea Samples are handled through step S1 to step S4 is composed
Area substitutes into the identification model and carries out corresponding qualification result, determines the quality of the Tea Samples.
Further, it is 80 mesh sample sifters that tea powder sieving is used in the step S1.
Further, the quality of every portion tea powder sample is 10-15g in the step S1.
Further, the near infrared spectrometer in the step S2 is Antaris II Fourier near infrared spectrometer.
Further, the acquisition parameter of the Antaris II Fourier near infrared spectrometer is as follows: scanning times 60
It is secondary, resolution ratio 8cm-1, spectral region 10000-4000cm-1。
Further, it is 9 that the spectrum of spectrum smoothing processing, which is smoothly counted, in the step S3.
Further, the particular content of the step S4 are as follows: by all band light of xls file format obtained in step S3
Modal data is divided into 30 subintervals and carries out random coded, wherein spectrum area is left out in 0 expression, 1 indicates to select spectrum area, is arranged in 50
Chiasma operator is iterated screening in conjunction with Partial Least Squares, is screened and is referred to using the maximum value of " R/W " parameter as wavelength
Mark show that optimized spectrum district's groups are closed, as the light spectral for participating in modeling.
Further, the parameter of the genetic algorithm is as follows: chromosome item number is 50, and gene number is 30, and variation is general
Rate is 0.01, crossover probability 0.4-0.6, and the number of iterations is 10-15 generation, and wherein R indicates that related coefficient, W indicate cross validation
Root-mean-square error.
Further, algorithm used in founding mathematical models is Partial Least Squares in the step S5.
Compared with the prior art, the invention has the following beneficial effects: the present invention carries out wavelength screening using genetic algorithm,
The model running time can be shortened, improve model stability, model is made to reach optimum state;The present invention in a few minutes can be simultaneously
Multiple samples are measured, high degree shortens the detection that mode and traditional chemical analysis method are manually evaluated in tea leaf quality identification
Time improves working efficiency;Using the tea judgement expert of profession according to national standard to the judge data of tea sample as model according to
According to ensure that this method qualification result is accurate, science;Be conducive to standardization of the tea market realization to tea leaf quality management.
Detailed description of the invention
Fig. 1 is the flow chart that the present invention establishes quality of Oolong tea identification model.
Fig. 2 is the correlativity figure of quality identification result of the present invention Yu practical review result.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and embodiments.
Fig. 1 is please referred to, the present invention provides a kind of identification method for oolong quality based on near-infrared spectrum technique, feature
Be the following steps are included:
Step S1: the preparation of tea powder sample: the Tea Samples in each oolong tea place of production are collected, by the Tea Samples through crushing
Machine is ground into tea powder, and the tea powder crosses 80 mesh sample sifters, and is numbered after taking the lower 10-15g tea powder of sieve, and encapsulation refrigeration is used as tea
Powder sample;
Step S2: the acquisition of spectral information: the near infrared spectrum information collection work based on Workflow setting tea powder sample
It flows, acquires the tea powder sample in diffusing near infrared range using Antaris II Fourier near infrared spectrometer
Spectrum information, acquisition parameter are as follows: scanning times are 60 times, resolution ratio 8cm-1, spectral region 10000-4000cm-1;
Step S3: the pretreatment of spectral information: collected to the step S2 unrestrained anti-using infrared processing software OPUS
It penetrates spectral information and carries out spectrum smoothing processing, second order derivation and normalized, while treated data are from spc spectrum format
Be converted to xls file format;It is 9 that the spectrum of spectrum smoothing processing, which is smoothly counted,;
Smoothing processing is for eliminating noise, and to filtering off, high-frequency noise is especially effective;Second order derivation can effectively eliminate baseline
And background interference, resolution ratio and sensitivity are improved, but it can amplify noise simultaneously, and signal-to-noise ratio is caused to reduce;Normalizing used herein
Changing processing is vector normalized, it is therefore an objective to eliminate the influence that change in optical path length generates spectrum.
Step S4: the wavelength screening based on genetic algorithm: by all band spectrum of xls file format obtained in step S3
Data are divided into 30 subintervals and carry out random coded, wherein 0 indicates to leave out spectrum area, 1 indicates to select spectrum area, contaminate in setting 50
Colour solid crossover operator is iterated screening in conjunction with Partial Least Squares, is screened and is referred to using the maximum value of " R/W " parameter as wavelength
Mark show that optimized spectrum district's groups are closed, as the light spectral for participating in modeling;The parameter of the genetic algorithm is as follows: chromosome item number
It is 50, gene number is 30, mutation probability 0.01, crossover probability 0.4-0.6, and the number of iterations is 10-15 generation, wherein R
Indicate that related coefficient, W indicate the root-mean-square error of cross validation;
By 10000-4000cm in one embodiment of the invention-1All band spectroscopic data be divided into 30 subintervals, i.e.,
Gene number be 30, (10000-4000)/30=200, i.e. each subinterval just have 200 data, to 30 subintervals into
Row random coded, wherein 0 indicates to leave out spectrum area, 1 indicates that spectrum area, such as (4000-4200) section is selected to be encoded to 0, (4200-
4400) section is encoded to 1, and (4400-4600) section is encoded to 1, and so on, coding is that computer is random, thus there is a string encoding.
Such as (011001101001100110100110011010).
50 kinds of chiasma operators are set, indicate there are 50 kinds of simulation nature species heredity sides during genetic algorithm
The cross and variation operator of formula indicates the dyeing of similar (011001101001100110100110011010) random coded of 50 strings
Body carries out the iteration screening that wavelength is carried out based on genetic algorithm combination Partial Least Squares.
Using the maximum value of " R/W " parameter as wavelength screening index, the available most accurately spectrum of result is filtered out
Combination, as the final light spectral for participating in modeling;Wherein R is the related coefficient (better closer to 1) of final result, and W is indicated
Root-mean-square error (better closer to 0).
If finally screen out is encoded to 100110100110011010011001101001
That is, first segment (4000-4200) section is encoded to 1 reservation, (4200-4400) section is encoded to 0 and leaves out, (4400-
4600) section be encoded to 0 leave out, (4600-4800) section etc. is encoded to 1 wave band and leaves, and so on, obtain optimal spectrum district's groups
It closes, thinks the light spectral for participating in modeling;
Step S5: the foundation of model is identified: according to " GB/T23776-2009 tealeaves organoleptic evaluation method " to each tealeaves sample
Product carry out quality identification, and the processing for carrying out step S1 to step S4 to every a Tea Samples obtains corresponding light spectral,
Qualification result is established using Matlab mathematical software and handles obtained light spectral and identifies model correspondingly.
Wherein tea judgement expert carries out oolong tea sample according to " GB/T23776-2009 tealeaves organoleptic evaluation method "
Sensory, the reference index of the model are as follows: coefficient R (Correlation Coefficient), cross validation root mean square
Error W(Root Mean Square Error of Calibration, RMSEC);In step S5 used in founding mathematical models
Algorithm is Partial Least Squares;
Step S6: the measurement of quality score: the spectrum that a Tea Samples are handled through step S1 to step S4 is composed
Area substitutes into the identification model and carries out corresponding qualification result, determines the quality of the Tea Samples.
Fig. 2 is the correlativity figure of quality identification result of the present invention Yu practical review result, is determined in figure with hundred-mark system
The comprehensive score of quality of Oolong tea, abscissa is the quality comprehensive score measured using method of the present invention in figure, and is indulged
Coordinate is that tea judgement expert comments oolong tea sample progress sense organ according to " GB/T23776-2009 tealeaves organoleptic evaluation method "
Resulting comprehensive score is examined, as can be seen from the figure the compatible degree of the two is very high, therefore method of the present invention has
Very high feasibility, and qualification process is more rapid.
The foregoing is merely presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with
Modification, is all covered by the present invention.
Claims (6)
1. a kind of identification method for oolong quality based on near-infrared spectrum technique, it is characterised in that the following steps are included:
Step S1: the preparation of tea powder sample: collecting the Tea Samples in each oolong tea place of production, and the Tea Samples are ground through pulverizer
Tea powder is worn into, the tea powder is used as tea powder sample after sieving, number and encapsulation refrigeration;
Step S2: the acquisition of spectral information: the near infrared spectrum information collection workflow based on Workflow setting tea powder sample,
The tea powder sample is acquired in the spectrum information that diffuses near infrared range using near infrared spectrometer;
Step S3: it the pretreatment of spectral information: diffuses using infrared processing software OPUS is collected to the step S2
Spectrum information carries out spectrum smoothing processing, second order derivation and normalized, while treated that data are converted from spc spectrum format
For xls file format;
Step S4: the wavelength screening based on genetic algorithm: by all band spectroscopic data of xls file format obtained in step S3
It is equally divided into multiple subintervals and carries out random coded, all various combinations in each subinterval constitute search space, " R/W " is used to join
Several maximum values filters out an optimized spectrum district's groups and closes, as the light spectral for participating in modeling as wavelength screening index;
Step S5: identify the foundation of model: according to " GB/T23776-2009 tealeaves organoleptic evaluation method " to each Tea Samples into
Row quality identification, and the processing for carrying out step S1 to step S4 to every a Tea Samples obtains corresponding light spectral, uses
Matlab mathematical software establishes qualification result and handles obtained light spectral identifies model correspondingly;
Step S6: the measurement of quality score: the light spectral generation that a Tea Samples are handled through step S1 to step S4
Enter the identification model and carry out corresponding qualification result, determines the quality of the Tea Samples;
The particular content of the step S4 are as follows: by all band spectroscopic data average mark of xls file format obtained in step S3
Random coded is carried out at 30 subintervals, wherein 0 indicates to leave out spectrum area, 1 indicates to select spectrum area, and 50 kinds of chiasmas are arranged
Operator is iterated screening in conjunction with Partial Least Squares, using the maximum value of " R/W " parameter as wavelength screening index, obtains most
Good spectrum district's groups are closed, as the light spectral for participating in modeling;
The parameter of the genetic algorithm is as follows: chromosome item number is 50, and gene number is 30, and mutation probability 0.01 intersects
Probability is 0.4-0.6, and the number of iterations is 10-15 generation, and wherein R indicates related coefficient, and better closer to 1, W indicates cross validation
Root-mean-square error;It is better closer to 0.
2. the identification method for oolong quality according to claim 1 based on near-infrared spectrum technique, it is characterised in that: institute
Stating tea powder in step S1 and being sieved used is 80 mesh sample sifters.
3. the identification method for oolong quality according to claim 1 based on near-infrared spectrum technique, it is characterised in that: institute
The quality for stating every portion tea powder sample in step S1 is 10-15g.
4. the identification method for oolong quality according to claim 1 based on near-infrared spectrum technique, it is characterised in that: institute
Stating the near infrared spectrometer in step S2 is Antaris II Fourier near infrared spectrometer.
5. the identification method for oolong quality according to claim 4 based on near-infrared spectrum technique, it is characterised in that: institute
The acquisition parameter for stating Antaris II Fourier near infrared spectrometer is as follows: scanning times are 60 times, resolution ratio 8cm-1, spectrum
Range is 10000-4000cm-1。
6. the identification method for oolong quality according to claim 1 based on near-infrared spectrum technique, it is characterised in that: institute
It is 9 that the spectrum for stating spectrum smoothing processing in step S3, which is smoothly counted,.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510343494.8A CN104897608B (en) | 2015-06-19 | 2015-06-19 | A kind of identification method for oolong quality based on near-infrared spectrum technique |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510343494.8A CN104897608B (en) | 2015-06-19 | 2015-06-19 | A kind of identification method for oolong quality based on near-infrared spectrum technique |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104897608A CN104897608A (en) | 2015-09-09 |
CN104897608B true CN104897608B (en) | 2019-05-03 |
Family
ID=54030410
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510343494.8A Active CN104897608B (en) | 2015-06-19 | 2015-06-19 | A kind of identification method for oolong quality based on near-infrared spectrum technique |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104897608B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105950749A (en) * | 2016-06-08 | 2016-09-21 | 福建农林大学 | Southern Fujian oolong tea identification method based on DNA barcode technology |
CN105938093A (en) * | 2016-06-08 | 2016-09-14 | 福建农林大学 | Oolong tea producing area discrimination method based on combination of genetic algorithm and support vector machine |
CN106560702A (en) * | 2016-10-20 | 2017-04-12 | 中国计量大学 | Wuyi rock tea production place identification method through combination of electronic tongue and chromatographic separation technology |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101419166A (en) * | 2008-11-18 | 2009-04-29 | 江苏大学 | Tea quality nondestructive detecting method and device based on near-infrared spectrum and machine vision technology |
CN101620180A (en) * | 2009-05-08 | 2010-01-06 | 合肥美亚光电技术有限责任公司 | Method for rapidly detecting tea quality through near infrared technology |
CN102305772A (en) * | 2011-07-29 | 2012-01-04 | 江苏大学 | Method for screening characteristic wavelength of near infrared spectrum features based on heredity kernel partial least square method |
CN103175815A (en) * | 2013-03-06 | 2013-06-26 | 浙江大学 | Multi-wavelength LED-induced fluorescence tea quality nondestructive testing method and device |
CN103743698A (en) * | 2013-12-20 | 2014-04-23 | 贵州省分析测试研究院 | Method for performing sensory evaluation on tea by adopting near infrared spectrum |
CN104568930A (en) * | 2014-12-18 | 2015-04-29 | 昆明理工大学 | Method for determining content of catechuic acid in tea and tea products |
-
2015
- 2015-06-19 CN CN201510343494.8A patent/CN104897608B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101419166A (en) * | 2008-11-18 | 2009-04-29 | 江苏大学 | Tea quality nondestructive detecting method and device based on near-infrared spectrum and machine vision technology |
CN101620180A (en) * | 2009-05-08 | 2010-01-06 | 合肥美亚光电技术有限责任公司 | Method for rapidly detecting tea quality through near infrared technology |
CN102305772A (en) * | 2011-07-29 | 2012-01-04 | 江苏大学 | Method for screening characteristic wavelength of near infrared spectrum features based on heredity kernel partial least square method |
CN103175815A (en) * | 2013-03-06 | 2013-06-26 | 浙江大学 | Multi-wavelength LED-induced fluorescence tea quality nondestructive testing method and device |
CN103743698A (en) * | 2013-12-20 | 2014-04-23 | 贵州省分析测试研究院 | Method for performing sensory evaluation on tea by adopting near infrared spectrum |
CN104568930A (en) * | 2014-12-18 | 2015-04-29 | 昆明理工大学 | Method for determining content of catechuic acid in tea and tea products |
Non-Patent Citations (2)
Title |
---|
Optimization of Informative Spectral Variables for the Quantification of EGCG in Green Tea Using Fourier Transform Near-Infrared (FT-NIR) Spectroscopy and Multivariate Calibration;ZHIMING GUO et al.;《APPLIED SPECTROSCOPY》;20111231;第65卷(第9期);第1062-1067页 |
基于近红外光谱法的大佛龙井茶品质评价研究;周小芬等;《光谱学与光谱分析》;20121130;第32卷(第11期);摘要、第2971页右栏第2段至第2973页右栏倒数第1段、表1-3、图1-3 |
Also Published As
Publication number | Publication date |
---|---|
CN104897608A (en) | 2015-09-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110763698B (en) | Hyperspectral citrus leaf disease identification method based on characteristic wavelength | |
CN108181263B (en) | Tobacco leaf position feature extraction and discrimination method based on near infrared spectrum | |
CN104965973B (en) | A kind of Apple Mould Core multiple-factor Non-Destructive Testing discrimination model and method for building up thereof | |
Ren et al. | Intelligent assessment of tea quality employing visible-near infrared spectra combined with a hybrid variable selection strategy | |
CN104897608B (en) | A kind of identification method for oolong quality based on near-infrared spectrum technique | |
CN103389281A (en) | Pu'er tea clustering analysis method based on near-infrared spectroscopy | |
CN105938093A (en) | Oolong tea producing area discrimination method based on combination of genetic algorithm and support vector machine | |
CN110132938B (en) | Characteristic data extraction method for identifying rice types by Raman spectroscopy | |
CN108844917A (en) | A kind of Near Infrared Spectroscopy Data Analysis based on significance tests and Partial Least Squares | |
CN106568738A (en) | Method of using near infrared spectroscopy to rapidly determine fresh leaves of tea in different quality grades | |
CN112669915B (en) | Pear nondestructive testing method based on neural network and near infrared spectrum | |
CN111257277B (en) | Tobacco leaf similarity judgment method based on near infrared spectrum technology | |
CN107796782A (en) | Redrying quality stability evaluation method based on tobacco leaf characteristic spectrum consistency metric | |
CN106248617A (en) | Based near infrared tobacco tar detection method | |
CN104062262A (en) | Crop seed variety authenticity identification method based on near infrared spectrum | |
CN109870421A (en) | It is a kind of based on visible light/near-infrared spectrum analysis incrementally timber varieties of trees classifying identification method | |
CN102735642A (en) | Method for quickly and losslessly identifying virgin olive oil and olive-residue oil | |
CN103743705A (en) | Rapid detection method for sorghum halepense and similar species | |
CN105138834A (en) | Tobacco chemical value quantifying method based on near-infrared spectrum wave number K-means clustering | |
CN108344701A (en) | Paraffin grade qualitative classification based on hyperspectral technique and quantitative homing method | |
CN102680427A (en) | Method for identifying surface fragrance quality of cigarettes by near infrared spectral analysis technology | |
CN105911000A (en) | Characteristic wave band based blood spot egg on-line detection method | |
CN107515204A (en) | Detection method using NIR to bean paste sweet tea valve fermenting-ripening degree | |
CN107886115A (en) | A kind of tealeaves mid-infrared light profile classification method of adaptively possible C mean clusters | |
EP3961189B1 (en) | Cigarette mainstream smoke spectral data sensory evaluation method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |