CN104122308A - Difference degree calculation method and tea quality identification method based on electronic tongue detection - Google Patents
Difference degree calculation method and tea quality identification method based on electronic tongue detection Download PDFInfo
- Publication number
- CN104122308A CN104122308A CN201410336153.3A CN201410336153A CN104122308A CN 104122308 A CN104122308 A CN 104122308A CN 201410336153 A CN201410336153 A CN 201410336153A CN 104122308 A CN104122308 A CN 104122308A
- Authority
- CN
- China
- Prior art keywords
- sample
- standard
- quality
- pca
- tea
- 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 invention belongs to the technical fields of application methodology and electronic tongue detection, and discloses a difference degree calculation method and tea quality identification method based on the electronic tongue detection. The difference degree calculation method comprises the steps of electronic tongue detection, PCA analysis, and difference degree calculation; wherein the difference degree calculation adopts a recently improved calculation mode, each coordinate in the PCA score matrix is taken into account during the calculation process in a form of distance between each coordinate and the center point so as to reduce the affect of the random errors during the experiment process, and an sample acceptable fluctuation range is formed at the same time. The tea quality identification is on the basis of the electronic tongue difference degree detection of tea samples, the same tea species is taken as the standard, the credibility interval is set according to the difference degree in each group, and thus established is a standard sample experiment measured value error range. If the measured value of a tea sample is in the credibility interval, the sample is qualified; and if the measured value of a tea sample is not in the credibility interval, the sample is disqualified. The method can be used to control tea quality, and is a more efficient and stable detection method for market, enterprises, and detection organizations.
Description
Technical field
The invention belongs to application process and learn and electronic tongue detection technique field, a series of application based on electronic tongue detection technique such as particularly tea leaf quality detection, tea grades classification, tea-drinking differentiation.
Background technology
Tea was the good wood of southern china originally, and tealeaves is as a kind of famous health beverages, and it is the contribution of ancient china south people's centering state cooking culture, is also the contributions of Chinese people to world's cooking culture.Tea tree leaf is made tealeaves, and soaked rear use has the effect of cardiac stimulant, diuresis.Tea tree planting 3 years just can picking leaves.The tender shoots that grows 4-5 leaf is plucked in general clear and bright front and back, and the tea quality of making of this tender shoots is very good, belongs to the treasure in tea.Tea and cocoa, the large non-alcoholic drink in the coffee Bing Cheng world today three, first of the large beverage in the world three.
It is generally acknowledged, tea quality is one group of process that inherent characteristic meets the demands of tea products, i.e. the inherent characteristic such as the edibility of tealeaves, health, trophism, security, credibility, economy and aesthetic property.
1. the assessment of tea quality
China's tealeaves type is many, and kind is complicated, is difficult to quantize quality standard, and except state compulsory standard, to tea quality, assessment does not have unified standard.China has set up tea quality test stone gradually, has had certain appraisement system, comprises that tealeaves overall assessment analysis index and tealeaves quantize evaluation index.
Tealeaves overall assessment is analyzed four kinds: 1. quality level index, 2. Customer Satisfaction Index, 3. mass loss index, 4. market degree of purification index.
2. tealeaves evaluation analysis index: 1. spot-check comprehensive qualified rate, 2. organoleptic quality qualification rate, 3. physical and chemical index qualification rate, 4. sanitary index qualification rate, 5. Proper Packing index.
The problems such as tea quality problem is mainly that pesticide residue, harmful heavy metal are residual, harmful microorganism, non-tea foreign matter and dust pollution.These all by force, do not notice that with current China part tealeaves practitioner quality safety consciousness hygienic issues in process etc. is relevant yet.Separately there is the illegal businessman of only a few, in tealeaves, add artificial color etc.Not only endangered the quality safety of tealeaves, returned whole tealeaves industry and bring negative effect.
The standard model of known its quality of take is reference, by calculating the difference degree detecting based on electronic tongue between standard model and target sample, to judge the quality discrepancy between target sample and standard model, and then reaches the object that target sample quality is differentiated.It is a kind of relatively more objective quality discrimination method.
The electrochemical properties of different liquids sample is different, by the electrochemical characteristic of tracer liquid, can effectively distinguish two kinds of different liquid.The main tool of the detection of the liquid electrochemical characteristic of evaluating and testing for sense organ is at present electronic tongue.Electronic tongue technology is a kind of by the artificial cognition technology of simulation human body taste perception mechanism, and it is conventionally by take the signal acquiring system that responsive cross-point sensor is array, and exciting with acquisition system and this three part-structure of multivariate statistical analysis algorithm of signal forms.
Electronic tongue is as a kind of testing tool of liquid electrochemical characteristic, and its principle is: solution to be checked is applied to significantly pulse signal of multifrequency, make to form electrostatic double layer between solution and sensor and discharge and recharge; Gather the current instantaneous value in solution in charge and discharge process simultaneously; The current value collecting is combined into the signal of reflection curent change, analyze afterwards the signal of test sample and standard with reference to difference degree between the signal of sample, with differentiation degree, (or be difference degree, with DI, represent) represent the size of difference degree, and then judge that whether test sample is identical with the quality of standard model.
According to the principle of electronic tongue, its general structure can be divided into three parts.Sensor array, in order to simulate biological taste perception system, is converted to electric signal by the physical chemistry signal of kinds of ingredients in solution; Signal acquiring system, it is that the signals collecting of perception is stored in computing machine, is equal to biological nervous system by cynapse the cerebripetal nerve center of taste bud neurocyte operating potential; Intelligence Multivariate Statistical Analysis System, it is exactly that simulation human brain utilizes the algorithm for pattern recognition in Chemical Measurement process, analyze and identify the electric signal of collection.
In electronic tongue testing process, between analytic sample during difference degree, the main data processing method using is principal component analysis (PCA)---PCA (Principal component analysis), it is a kind of by the technology to data analysis again of carrying out after dimensionality reduction to data, the method is ripe data processing method, follows identical computing formula.It can find out topmost composition from complicated data structure, removes noise and redundancy, thereby extracts main information hiding in data, farthest the relation between announcement and legacy data.Because its method is simple, printenv restriction, so be widely used in biological information science and computer graphics.
Generally before using PCA, first extract minutiae (flex point, peak value etc.) from signal, the current value that the unique point of usining is corresponding or have other value of direct corresponding relation to analyze the data set of use as PCA with this current value.PCA can obtain the principal component analysis (PCA) score matrix that detects sample after analyzing, it (is the score value that obtains that every a line is deposited one-time detection that matrix is sequentially arranged according to detection, by principal ingredient 1 score value, principal ingredient 2 score value ..., be discharged to last composition score value).
In the prior art, the computing method of difference degree have multiple, the most frequently used method is, according to the score value that obtains that obtains score value and principal ingredient 2 of the principal ingredient 1 in score matrix, to each detection make scatter diagram (with principal ingredient 1 must be divided into X coordinate figure, with principal ingredient 2, must be divided into Y coordinate figure), by loose some line of same sample, form the sample area of this sample, finally according to the area ratio between the sample area of different samples, obtain the differentiation degree DI value between different samples.
Adopt said method to calculate difference degree, cannot consider the impact of the point that is positioned at sample area inside, accuracy is not high.In addition, not yet have at present electronic tongue is detected to the research of differentiating for tea leaf quality.
Summary of the invention
The object of this method is, for above-mentioned deficiency, providing a kind of can all take the impact of each principal ingredient score into account, difference degree computing method.And this new method is applied to the Quality Identification of tealeaves.
For achieving the above object, first the present invention discloses a kind of sample difference degree computing method that detect based on electronic tongue, comprises electronic tongue detecting step, PCA analytical procedure, difference degree calculation procedure; It is characterized in that, described difference degree calculation procedure is specific as follows:
Major component 1 in the score matrix analyze obtaining with PCA respectively and major component 2 score value as X coordinate and Y coordinate, the PCA analysis result to two kinds of samples in the same coordinate system carries out mark, calculates the difference degree of two kinds of sample rooms by formula ();
In formula, D
ddistance value for sample i and sample R; Δ mv is the Euclidean distance at sample i center and sample R center; δ
rfor each measurement data (coordinate of point) of sample R standard deviation on center position; δ
ifor each measurement data of sample i standard deviation in population on center position;
Said method, by each of PCA score matrix to coordinate to consider in computation process to the form of central point distance, can well reduce the impact of accidental error in experimentation, also can form the fluctuation range of accepting of sample simultaneously.
Based on said method, the present invention further proposes a kind of tea leaf quality discrimination method detecting based on electronic tongue, comprises the steps:
Step 1, brewed standard model, is divided into two groups of A, B, and every group of n part is total to 2n part standard model by electronic tongue to two groups and detects, and carries out PCA analysis to detecting the 2n group raw data obtaining, and obtains 2n PCA score matrix; N the PCA score matrix that detects the acquisition of A group formed to formation n to PCA score matrix one to one with n the PCA score matrix that detection B group obtains.
For every pair of PCA score matrix, respectively with the major component 1 in PCA score matrix and major component 2 score value as X coordinate and Y coordinate, PCA analysis result to two kinds of samples in the same coordinate system carries out mark, calculates the difference degree of two minutes sample rooms by formula ().
In formula, D
ddistance value for sample i and sample R; Δ mv is the Euclidean distance at sample i center and sample R center; δ
rfor each measurement data (coordinate of point) of sample R standard deviation on center position; δ
ifor each measurement data of sample i standard deviation in population on center position.
N the difference degree value obtaining calculated in combination to PCA score matrix based on n, form standard tealeaves quality differential degree interval.
Step 2, the standard tealeaves quality differential degree obtaining based on step 1 is interval, according to formula (two) Criterion sample quality control line;
Wherein, U
lfor the quality control line of standard model with test sample;
mean value for the difference number of degrees value of each measurement data; T (a, f) is is that a fiducial interval is the critical value of the t of the t-method of inspection under f in degree of freedom; N is number of degrees of freedom; S is the standard deviation of sample.
Step 3, gets respectively n part standard model and n part target sample, obtains target sample interval with respect to the quality differential degree of standard model with reference to step 1.
Step 4, contrast standard tea leaf quality control line and target sample are interval with respect to the quality differential degree of standard model, target sample is with respect to the quality differential degree interval of standard model lower than standard tealeaves quality control line, and target sample is identical with standard model quality.
As preferably, in said method, the detection of sample adopts cross detection method, for example, every portion of A, two groups of samples of B is marked as A1, A2, A3 ... An and B1, B2, B3 ... Bn, detects cross detection method and refers to by electronic tongue equipment 2n duplicate samples by A1, B1, A2, B2, A3, B3 ... the order of An, Bn detects.
As a kind of improvement, the present invention proposes the another kind of tea leaf quality discrimination method detecting based on electronic tongue, comprises the steps:
Step 1, brewed standard model, is divided into two groups of A, B, and every group of n part is total to 2n part standard model by electronic tongue to two groups and detects, and carries out PCA analysis to detecting the 2n group raw data obtaining, and obtains 2n PCA score matrix; N the PCA score matrix that detects the acquisition of A group formed to formation n to PCA score matrix one to one with n the PCA score matrix that detection B group obtains.
For every pair of PCA score matrix, respectively with the major component 1 in PCA score matrix and major component 2 score value as X coordinate and Y coordinate, PCA analysis result to two kinds of samples in the same coordinate system carries out mark, calculates the difference degree of two minutes sample rooms by formula ().
In formula, D
ddistance value for sample i and sample R; Δ mv is the Euclidean distance at sample i center and sample R center; δ
rfor each measurement data (coordinate of point) of sample R standard deviation on center position; δ
ifor each measurement data of sample i standard deviation in population on center position.
N the difference degree value obtaining calculated in combination to PCA score matrix based on n, form standard tealeaves quality differential degree interval.
Step 2, the standard tealeaves quality differential degree obtaining based on step 1 is interval, according to formula (two) Criterion sample quality control line;
Wherein, U
lfor the quality control line of standard model with test sample;
mean value for the difference number of degrees value of each measurement data; T (a, f) is is that a fiducial interval is the critical value of the t of the t-method of inspection under f in degree of freedom; N is number of degrees of freedom; S is the standard deviation of sample.
Step 3, gets respectively n part standard model and n part target sample, obtains target sample interval with respect to the quality differential degree of standard model with reference to step 1.
Step 4, the target sample obtaining based on step 3 is interval with respect to the quality differential degree of standard model, according to formula (two), calculates target sample with respect to the product control control value of standard model.
Step 5, contrast standard tea leaf quality control line and target sample are with respect to the product control reference point of standard model, and the latter is less than or equal to the former, and target sample is identical with standard model quality.
The tea leaf quality discrimination method that the present invention proposes, standard model is divided into groups to detect contrast, the detection permissibl e skew range obtaining (equipment Inspection error and manual operation error), and set up accordingly a quality control line, when differentiating target sample, if difference degree interval is in this permissibl e skew range, think that target sample is identical with standard model quality, the method can be for product acceptance test, quality discriminating etc. the method utilizes quality control line to distinguish judgement to the quality of object sample, thereby reach the object of quality control, method simple, intuitive, strong operability.
The electronic tongue difference degree that said method is based on Tea Samples detects, with tealeaves of the same race as standard, in its group, difference degree arranges credibility interval, thereby can Criterion sample experiment measuring value scope, be convenient to judgement detection sample quality and whether reach in the fiducial interval of standard model place, and then determine the quality that detects sample.The foundation of this application method, will be beneficial to the control of tea leaf quality, by for market, enterprise, testing agency provide more efficient, stable detection method, is convenient to tea grades classification, tealeaves discriminating, tea-drinking exploitation, tea-drinking quality control etc.
Accompanying drawing explanation
Fig. 1 is the difference degree computing method schematic flow sheet detecting based on electronic tongue of the present invention.
Fig. 2 be in standard model group and standard model and target sample between the graphical schematic diagram of difference degree.
Fig. 3 is standard model A quality control line schematic diagram.
Embodiment
Below in conjunction with specific embodiment, content of the present invention is further described.
So the invention belongs to and a kind ofly take tea leaf quality difference degree interval as basic electronic tongue application process, comprise electronic tongue detection, analysis of experimental data of brewed, the tea of tea etc.Its feature is, data analysis is by the difference degree of standard tealeaves sample room, and in conjunction with t check, Criterion tea leaf quality difference degree is interval, and recycling quality control line judges tea leaf quality.
What reacted in standard tealeaves quality differential degree interval is the inevitable error occurring in the processes such as tealeaves is brewed, detection.Because inevitable error, even twice detection of same sample, its result is also different, therefore, the present invention by same sample not homogeneous testing result calculate difference degree, find out between the generating region of this error, as quality, differentiate delegatable acceptability limit.
The interval foundation of difference degree of the present invention be take electronic tongue experimental data as basis.Use electronic tongue to carrying out cross detection between standard model, the quality differential degree that the data that obtain are calculated formation standard tealeaves through difference degree formula is interval.Afterwards object Tea Samples and standard model are carried out to cross detection, utilize difference degree formula to calculate the difference degree between object sample and standard model.Finally by this difference degree of quality control line computation whether in the quality differential degree interval of standard tealeaves, and then judge whether the quality of object tealeaves reaches standard.
Below for an instantiation of tealeaves product control method of the present invention.
1, tealeaves chooses
Standard tealeaves is chosen: standard tealeaves by the tealeaves detecting, quality is qualified, is usingd it as the standard of setting up tealeaves difference degree interval for.Standard tealeaves should be the same place of production with the ad eundem of company's made with kind tealeaves.In this test, the one-level West Lake Dragon Well tea that standard model selects Hangzhou Tea Co., Ltd of HTC to produce, the secondary West Lake Dragon Well tea that object sample Wei Mou company produces.
2, tea is brewed
The brewed of tea carried out according to GB GB/T23776-2009.(green tea) tea-water proportion is 1:50,85 ℃ of water temperatures, 5 minutes brewed time.After first putting tealeaves, add water, tea filters through 80 mesh sieves afterwards, is cooled to room temperature, waits to be detected.
3, the foundation of tea quality differential degree interval and product nature controlling line
This experiment checkout equipment adopts the electronic tongue equipment of intelligence sense organ seminar of Zhejiang Prov Industrial And Commercial University research and development, data analyzing and processing software version 1.0, software copyright registration number: 0517194.
3.1, electronic tongue sensor is processed intelligence tongue system
Six experimental datas that electrode 1Hz, 10Hz, tri-frequency bands of 100Hz gather of electronic tongue platinum palladium tungsten titanium silver of using.
On electronic tongue sensor surface, be coated with and spread grinding paster, be placed on polishing paper and polish to smooth surface.Wipe remained on surface grinding paster away, sensor is placed in to acetone, ethanol, ultrapure water successively, each ultrasonic cleaning 15min.
3.2, the electronic tongue of standard Tea Samples detects
Sensor after processing is placed in to standard model liquid, and connecting electronic tongue, sensor, computer, open electronic tongue power supply, operation electronic tongue software, and electronic tongue parameter is set to: starting potential 1V, final voltage-1V, stepped voltage 0.2V; S1-S6 transducer sensitivity is 10
-4.Select required sensor (S1-S6) and transducer sensitivity (10
-4), click " sensor preheating " on software, electronic tongue is carried out to thermal pretreatment.After preheating, sensor is rinsed with ultrapure water, be placed in afterwards " cleaning sensor " in ultrapure water, clicked on software sensor is cleaned.
Sensor is wiped away dry being placed in the standard model liquid that reaches room temperature after cleaning, and clicks " starting scanning " and carries out sample detection, image data.
Standard model detect to adopt cross detection, same tealeaves be divided into two groups brewed, detect and hocket.Every group is detected taking-up 15ml sample at every turn, is placed in electronic tongue special glass cup.After each detection, all to carry out ultrapure water flushing and use " cleaning sensor " order on software to clean sensor, wipe away afterwards stem grafting and the detection of carrying out next sample.Electronic tongue is used concrete steps as follows:
A. electronic tongue is connected with computer, opens electronic tongue power supply, the operation supporting data acquisition of electronic tongue and analysis software (smartongue), click " hardware check " in " system setting " on software interface, carries out hardware check.
B. hardware detection normal after, click " directly image data " in " image data " on software interface, eject new interface: Sampling_Figure.
C. in the right side of interface " Sampling_Figure " " Parameter ", can carry out the setting of electronic tongue detected parameters: " the Decreased Amplitude " in " Actuator " hurdle selects stepped voltage; In " Sensor Array " hurdle, choose required sensor, the sensitivity of drop-down selection respective sensor; In " Files Directory " hurdle, choose newly-built or directly select data storing path and filename; In " Samples Name " hurdle, fill in sample number into spectrum.
D. use ultrapure water to rinse sensor, and wipe away dry; Sensor is placed in to ultrapure water, clicks " cleaning sensor " on " Sampling_Figure " interface, carry out sensor and automatically clean.
E. after automatically having cleaned, sensor is taken out, with ultrapure water, rinse sensor, wipe away dry.
F. the sensor of wiping away after doing is placed in sample, clicks " pre-thermal sensor " in " Sampling_Figure " interface, carries out sensor preheating.
G. after preheating completes, repeating step d, e, be placed in sensor in sample afterwards, clicks " starting scanning " on " Sampling_Figure " interface, and fill in sample number into spectrum in " Samples Name " hurdle.
H., after having scanned, click " save data " on " Sampling_Figure " interface.
I. after having preserved, repeating step d, e, then carry out the detection of next sample, until all sample detection is complete.
J. sample detection complete after, reply sensor cleans (repeating step d, e at least 3 times), sensor is placed in is equipped with in saturated potassium chloride pullover afterwards.
Electronic tongue detection method adopts sample cross detection, and sample room alternately detects mutually, as sample has A, B, detection method be A1 → B1 → A2 → B2 → ... → An → Bn.
3.3, the difference degree of standard Tea Samples is interval calculates
Electronic tongue detects data analysis and uses software kit " matlab ".Data analysis step is as follows:
A. import program code: the file of the catalogue of matlab being selected to code place.
B. import data: the raw data after electronic tongue scanning can be kept in the file of " mdb " form, by the data importing of this file " matlab ", after this carries out the computing of difference degree.
C. record the results list: the value saving result that obtains difference degree.
D. complete after preservation, close each window interface and software.
3.4, standard Tea Samples matter control line is set up
With reference to Fig. 1, first electronic tongue is detected to data and carry out PCA (principal component analysis (PCA)) differentiation, obtain PCA score matrix, using the first row (PC1) in matrix and the second row (PC2) respectively as x coordinate and y coordinate, the coordinate figure of each point of mark in same coordinate system, the D between calculation sample R and sample i
dvalue.Finally, set up D
dthe relation of value and sample room.
The difference degree formula that uses is:
Wherein,
D
ddistance value for sample i and standard model R;
Δ mv is the Euclidean distance at sample i center and sample R center;
δ
rfor each measurement data of standard model R standard deviation on center position;
δ
ifor each measurement data of sample i standard deviation in population on center position.
As shown in Figure 2, in figure, left-half is the difference degree schematic diagram in standard model a, and right half part is the difference degree schematic diagram of standard model a and object sample b.
Quality control line is the method for utilizing crosscheck, by the difference degree between repeated detection standard model and object sample, then in conjunction with t check, thereby the separatrix of settle the standard sample and object sample.
Quality control line computation formula is:
Wherein, U
lquality control line numerical value for standard model;
for the electronic tongue difference degree distance value of each check of standard model is rejected the mean value after maximin; N is the number of times of parallel testing; T is confidence factor (t can be found by t value table), and a is level of significance, and f is degree of freedom, f=n-2; S is the standard deviation of sample.
Two standard sets sample electronic tongue is detected to raw data and import in " Matlab ", use according to difference degree formula
the M file of writing calculates the electronic tongue difference degree between standard Tea Samples A, and then calculates standard tealeaves quality control line.The electronic tongue difference degree of Tea Samples B to be measured and standard Tea Samples A is also like this.
The electronic tongue difference degree distance value of testing for each 10 times between difference degree and standard model A and testing sample B in standard model A group is recorded as following table:
4, standard tealeaves quality control line is set up and interpretation of result
According to electronic tongue difference degree distance value and the quality control line numerical computational formulas of resulting standard tealeaves AA above:
can calculate standard tealeaves A quality control line numerical value, be specially: remove the maximin of the resulting difference degree of AA10 parallel laboratory test of numbering, all the other 8 values are tried to achieve average
standard deviation s=0.04297, gets β=0.05, f=n-2=6, and t (β, f)=2.31 of tabling look-up to obtain, substitution formula obtains U
l=0.8193, the quality control line numerical value of standard tealeaves A is 0.8193.Take sample test order as horizontal ordinate, and difference number of degrees value is ordinate, makes pattern analysis.Below quality control line, meet product quality requirement.
As shown in Figure 3, the standard model A quality control line of setting up by difference number of degrees value can obviously separate one-level Longjing tea A and secondary Longjing tea B, only has the electronic tongue difference number of degrees value of 1 testing sample B under quality control line, the difference number of degrees value of only having 10% testing sample is the requirement that meets difference number of degrees value between standard model, be less than 80%, therefore testing sample tealeaves B is the quality requirements that does not meet standard model tealeaves A, they can be classified, with this quality control line, also can be used as the normative reference of the quality control of the one-level West Lake Dragon Well tea that certain company produces.
Claims (7)
1. sample difference degree computing method that detect based on electronic tongue, is characterized in that, comprise electronic tongue detecting step, PCA analytical procedure, difference degree calculation procedure; It is characterized in that, described difference degree calculation procedure is specially: the major component 1 in the score matrix analyze obtaining with PCA respectively and major component 2 score value as X coordinate and Y coordinate, PCA analysis result to two kinds of samples in the same coordinate system carries out mark, calculates the difference degree of two kinds of sample rooms by formula ();
In formula, D
ddistance value for sample i and sample R; Δ mv is the Euclidean distance at sample i center and sample R center; δ
rfor each measurement data (coordinate of point) of sample R standard deviation on center position; δ
ifor each measurement data of sample i standard deviation in population on center position.
2. the tea leaf quality discrimination method detecting based on electronic tongue, is characterized in that, comprises the steps:
Step 1, by standard model, is divided into two groups of A, B, and after difference is brewed, every component becomes n part, is total to 2n part standard model and detects, and carry out PCA analysis to detecting the 2n group raw data obtaining by electronic tongue to two groups, obtains 2n PCA score matrix; N the PCA score matrix that detects the acquisition of A group formed to formation n to PCA score matrix one to one with n the PCA score matrix that detection B group obtains;
For every pair of PCA score matrix, respectively with the major component 1 in PCA score matrix and major component 2 score value as X coordinate and Y coordinate, PCA analysis result to two kinds of samples in the same coordinate system carries out mark, by formula (), calculates the difference degree between two duplicate samples;
In formula, D
ddistance value for sample i and sample R; Δ mv is the Euclidean distance at sample i center and sample R center; δ
rfor each measurement data (coordinate of point) of sample R standard deviation on center position; δ
ifor each measurement data of sample i standard deviation in population on center position;
N the difference degree value obtaining calculated in combination to PCA score matrix based on n, form standard tealeaves quality differential degree interval;
Step 2, the standard tealeaves quality differential degree obtaining based on step 1 is interval, according to formula (two) Criterion sample quality control line;
Wherein, U
lfor the quality control line of standard model with test sample;
mean value for the difference number of degrees value of each measurement data; T (a, f) is is that a fiducial interval is the critical value of the t of the t-method of inspection under f in degree of freedom; N is number of degrees of freedom; S is the standard deviation of sample;
Step 3, gets respectively n part standard model and n part target sample, obtains target sample interval with respect to the quality differential degree of standard model with reference to step 1;
Step 4, contrast standard tea leaf quality control line and target sample are interval with respect to the quality differential degree of standard model, target sample is with respect to the quality differential degree interval of standard model lower than standard tealeaves quality control line, and target sample is identical with standard model quality.
3. tea leaf quality discrimination method according to claim 2, is characterized in that, the electronic tongue of sample detects the cross detection method that adopts.
4. tea leaf quality discrimination method according to claim 2, is characterized in that, the brewed of tealeaves is 1:50 according to tea-water proportion, brewed 5 minutes of 85 ℃ of water temperature constant temperature, and tea filters through 80 mesh sieves afterwards, is cooled to room temperature, waits to be detected.
5. the tea leaf quality discrimination method detecting based on electronic tongue, is characterized in that, comprises the steps:
Step 1, brewed standard model, is divided into two groups of A, B, and every group of n part is total to 2n part standard model by electronic tongue to two groups and detects, and carries out PCA analysis to detecting the 2n group raw data obtaining, and obtains 2n PCA score matrix; N the PCA score matrix that detects the acquisition of A group formed to formation n to PCA score matrix one to one with n the PCA score matrix that detection B group obtains;
For every pair of PCA score matrix, respectively with the major component 1 in PCA score matrix and major component 2 score value as X coordinate and Y coordinate, PCA analysis result to two kinds of samples in the same coordinate system carries out mark, calculates the difference degree of two minutes sample rooms by formula ();
In formula, D
ddistance value for sample i and sample R; Δ mv is the Euclidean distance at sample i center and sample R center; δ
rfor each measurement data (coordinate of point) of sample R standard deviation on center position; δ
ifor each measurement data of sample i standard deviation in population on center position;
N the difference degree value obtaining calculated in combination to PCA score matrix based on n, form standard tealeaves quality differential degree interval;
Step 2, the standard tealeaves quality differential degree obtaining based on step 1 is interval, according to formula (two) Criterion sample quality control line;
Wherein, U
lfor the quality control line of standard model with test sample;
mean value for the difference number of degrees value of each measurement data; T (a, f) is is that a fiducial interval is the critical value of the t of the t-method of inspection under f in degree of freedom; N is number of degrees of freedom; S is the standard deviation of sample;
Step 3, gets respectively n part standard model and n part target sample, obtains target sample interval with respect to the quality differential degree of standard model with reference to step 1;
Step 4, the target sample obtaining based on step 3 is interval with respect to the quality differential degree of standard model, according to formula (two), calculates target sample with respect to the product control control value of standard model;
Step 5, contrast standard tea leaf quality control line and target sample are with respect to the product control reference point of standard model, and the latter is less than or equal to the former, and target sample is identical with standard model quality.
6. tea leaf quality discrimination method according to claim 5, is characterized in that, the electronic tongue of sample detects the cross detection method that adopts.
7. tea leaf quality discrimination method according to claim 5, is characterized in that, the brewed of tealeaves is 1:50 according to tea-water proportion, brewed 5 minutes of 85 ℃ of water temperature constant temperature, and tea filters through 80 mesh sieves afterwards, is cooled to room temperature, waits to be detected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410336153.3A CN104122308A (en) | 2014-07-15 | 2014-07-15 | Difference degree calculation method and tea quality identification method based on electronic tongue detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410336153.3A CN104122308A (en) | 2014-07-15 | 2014-07-15 | Difference degree calculation method and tea quality identification method based on electronic tongue detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104122308A true CN104122308A (en) | 2014-10-29 |
Family
ID=51767802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410336153.3A Pending CN104122308A (en) | 2014-07-15 | 2014-07-15 | Difference degree calculation method and tea quality identification method based on electronic tongue detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104122308A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105527328A (en) * | 2015-12-02 | 2016-04-27 | 浙江树人大学 | Construction method and application of electrochemical fingerprint of tea |
| CN105866203A (en) * | 2016-03-29 | 2016-08-17 | 福建农林大学 | Electronic tongue detection information-based Baihao Yinzhen taste grade discrimination method |
| CN105866223A (en) * | 2016-03-29 | 2016-08-17 | 福建农林大学 | Electronic tongue detection information-based Wuyi rock tea taste grade discrimination method |
| CN105866209A (en) * | 2016-03-28 | 2016-08-17 | 齐鲁工业大学 | Electronic tongue-based soybean protein isolate identification method |
| CN106353474A (en) * | 2016-10-10 | 2017-01-25 | 丽水学院 | Tea water distinguishing device and distinguishing method |
| CN107991444A (en) * | 2017-11-28 | 2018-05-04 | 宁德师范学院 | A kind of method and device for tea leaf quality detection |
| CN108108863A (en) * | 2016-11-25 | 2018-06-01 | 上海昆涞生物科技有限公司 | Laboratory system allowable range of error appraisal procedure based on Quality Control data |
| CN108717078A (en) * | 2018-05-28 | 2018-10-30 | 安徽农业大学 | A kind of teas method of discrimination based on chemical composition |
| CN109884257A (en) * | 2019-03-28 | 2019-06-14 | 南京林业大学 | The discrimination method of cyclocarya paliurus tea |
| CN110646546A (en) * | 2019-10-10 | 2020-01-03 | 中国农业科学院农产品加工研究所 | Meat product flavor compound profile analysis method |
| CN114994150A (en) * | 2022-05-31 | 2022-09-02 | 中国标准化研究院 | Electronic tongue rapid classification method for spicy degree of zanthoxylum schinifolium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101358958A (en) * | 2008-09-05 | 2009-02-04 | 上海海洋大学 | Shelf life forecasting model for hairtail |
| CN101975845A (en) * | 2010-09-21 | 2011-02-16 | 成都中医药大学 | Automatic detecting system for quality of traditional Chinese medicines |
| CN102053109A (en) * | 2009-10-30 | 2011-05-11 | 浙江工商大学 | Electronic tongue sensor array and method for controlling quality of beverage products |
-
2014
- 2014-07-15 CN CN201410336153.3A patent/CN104122308A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101358958A (en) * | 2008-09-05 | 2009-02-04 | 上海海洋大学 | Shelf life forecasting model for hairtail |
| CN102053109A (en) * | 2009-10-30 | 2011-05-11 | 浙江工商大学 | Electronic tongue sensor array and method for controlling quality of beverage products |
| CN101975845A (en) * | 2010-09-21 | 2011-02-16 | 成都中医药大学 | Automatic detecting system for quality of traditional Chinese medicines |
Non-Patent Citations (2)
| Title |
|---|
| GUANGYING ZHAO等: "Use of the smart tongue to monitor mold growth and discriminate between four mold species grown in liquid media", 《ANALYTICA CHIMICA ACTA》 * |
| 张爱霞: "基于差别度检验的电子舌应用方法学研究", 《中国博士学位论文全文数据库 工程科技I辑》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105527328A (en) * | 2015-12-02 | 2016-04-27 | 浙江树人大学 | Construction method and application of electrochemical fingerprint of tea |
| CN105866209B (en) * | 2016-03-28 | 2018-06-12 | 齐鲁工业大学 | A kind of soybean protein isolate discrimination method based on electronic tongues |
| CN105866209A (en) * | 2016-03-28 | 2016-08-17 | 齐鲁工业大学 | Electronic tongue-based soybean protein isolate identification method |
| CN105866203A (en) * | 2016-03-29 | 2016-08-17 | 福建农林大学 | Electronic tongue detection information-based Baihao Yinzhen taste grade discrimination method |
| CN105866223A (en) * | 2016-03-29 | 2016-08-17 | 福建农林大学 | Electronic tongue detection information-based Wuyi rock tea taste grade discrimination method |
| CN105866203B (en) * | 2016-03-29 | 2018-06-29 | 福建农林大学 | A kind of silver tip pekoe flavour Grade Judgment based on electronic tongues detection information |
| CN105866223B (en) * | 2016-03-29 | 2018-06-29 | 福建农林大学 | A kind of Wuyi cliff tea flavour Grade Judgment based on electronic tongues detection information |
| CN106353474A (en) * | 2016-10-10 | 2017-01-25 | 丽水学院 | Tea water distinguishing device and distinguishing method |
| CN106353474B (en) * | 2016-10-10 | 2018-07-03 | 丽水学院 | Tea identification device and discrimination method |
| CN108108863A (en) * | 2016-11-25 | 2018-06-01 | 上海昆涞生物科技有限公司 | Laboratory system allowable range of error appraisal procedure based on Quality Control data |
| CN108108863B (en) * | 2016-11-25 | 2021-12-14 | 上海昆涞生物科技有限公司 | Laboratory system error allowable range evaluation method based on quality control data |
| CN107991444A (en) * | 2017-11-28 | 2018-05-04 | 宁德师范学院 | A kind of method and device for tea leaf quality detection |
| CN108717078A (en) * | 2018-05-28 | 2018-10-30 | 安徽农业大学 | A kind of teas method of discrimination based on chemical composition |
| CN109884257A (en) * | 2019-03-28 | 2019-06-14 | 南京林业大学 | The discrimination method of cyclocarya paliurus tea |
| CN110646546A (en) * | 2019-10-10 | 2020-01-03 | 中国农业科学院农产品加工研究所 | Meat product flavor compound profile analysis method |
| CN114994150A (en) * | 2022-05-31 | 2022-09-02 | 中国标准化研究院 | Electronic tongue rapid classification method for spicy degree of zanthoxylum schinifolium |
| CN114994150B (en) * | 2022-05-31 | 2023-10-27 | 中国标准化研究院 | Electronic tongue rapid classification method for redpepper tingling degree |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104122308A (en) | Difference degree calculation method and tea quality identification method based on electronic tongue detection | |
| Hamilton et al. | Trade liberalisation, structural adjustment and intra-industry trade: a note | |
| CN101692053B (en) | Multi-sensing information fusion based instrumental intelligent evaluation method for quality of famous tea | |
| CN108037256B (en) | The rapid assay methods of rice eating-quality | |
| CN108663339A (en) | Corn online test method of going mouldy based on spectrum and image information fusion | |
| CN103389281A (en) | Pu'er tea clustering analysis method based on near-infrared spectroscopy | |
| CN103163217A (en) | Sound surface wave resonator series detection and detection method | |
| CN104713849A (en) | Method for quickly predicting tobacco mildew based on near-infrared spectrum analysis technique | |
| CN102692488A (en) | Jinhua ham grading and identifying method based on electronic nose technology | |
| CN107045013A (en) | Beef taste quality automatic classification detector and its detection method | |
| Wei et al. | The evaluation of sugar content and firmness of non-climacteric pears based on voltammetric electronic tongue | |
| CN103412047A (en) | Method for identifying true and false metal by ultrasonic non-destructive detection method | |
| CN107741439A (en) | A kind of evaluation method of the mellow and full property of beer wine body | |
| CN101685092A (en) | Method for judging alcoholization quality of flue-cured tobacco by pH detection value of tobacco leaf | |
| CN106970191A (en) | A kind of method of use intelligence tongue system to white spirit original wine rapid classification | |
| CN112033911A (en) | Method for rapidly identifying grade of tea based on chromatic aberration and ultraviolet spectrum | |
| Zhong et al. | Correlation analysis between color parameters and sensory characteristics of rice with different milling degrees | |
| CN105866223B (en) | A kind of Wuyi cliff tea flavour Grade Judgment based on electronic tongues detection information | |
| Xiao et al. | Discrimination of Xihulongjing tea grade using an electronic tongue | |
| CN116823363B (en) | Dish-washing machine performance rating system and evaluation method | |
| CN103487384B (en) | Based on the packed paddy composition method for quick of near infrared spectrum | |
| Lu et al. | Determination of rice sensory quality with similarity analysis-artificial neural network method in electronic tongue system | |
| CN106770477A (en) | One kind optimization sensing data and pattern-recognition differentiate and adulterated fast detecting method to nectar source | |
| CN103399012A (en) | Colorimetric sensing paste film and preparation method thereof | |
| CN105758897A (en) | White peony fragrance grade discriminating method based on electronic nose detection information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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: 20141029 |
|
| WD01 | Invention patent application deemed withdrawn after publication |