Background technology
The mankind obtain food, not only meet psychological need, also meet psychological needs and spiritual enjoyment.There is food that is good or peculiar flavour, except directly increasing appetite, promote digestive juice secretion, strengthen the function digested and assimilated except, people also can be made on sense organ to obtain happiness, the glamour place of Here it is flavour of food products.Flavour of food products refers to sensory stimulations such as the eye of human body, nose, tongues before and after food entrance, thus causes people to the comprehensive impression of its general characteristic, is in particular in the many aspects such as color, shape.For a long time, the inspection of flavour of food products and judge rely on artificial sense inspection usually, but artificial sense inspection has some limitations itself: the 1) process of artificial sense inspection needs prolonged exercise and experience accumulation, be generally rely on well-trained expert to come, ordinary consumer does not possess this ability; 2) artificial sense inspection is easily subject to the interference of external factor, even well-trained expert, their sensory sensitivity also can be subject to the interference of the factors such as sex, experience, the state of mind, health, motivation attitude and territorial environment, has influence on accuracy and the consistance of evaluation result.Therefore, invent a kind of intelligentized sensory analysis method, with indirect labor's organoleptic examination, correct artificial sense inspection deviation, improve objectivity and the consistance of sensory analysis result, have extremely important meaning to instructing food production, ensureing food quality and increasing food added value etc.
In recent years, along with the development of computing machine, microelectronics and material science, the Novel bionic sensor technologies such as vision, sense of smell and the sense of taste are come out one after another, and they obtain application in food quality intelligent Evaluation.Food product packets is containing the multiple organoleptic quality index such as color, shape, the quantity of information of single-sensor technical limit spacing is comprehensive not, finally have influence on the accuracy of testing result, in recent years, multi-sensor information fusion technology have also been obtained Preliminary Applications in food quality intelligent Evaluation, compared with single detection means, and the application of Multi-information acquisition in Intelligent Food organoleptic examination, the information that detection can be made to acquire more comprehensively, result is more objective, and similar with human perception process.
Can find out from the literature search such as paper and patent, at present in food quality intelligent Evaluation, biomimetic sensor adopts business-like instrument mostly, only play the effect of analytical instrument, detection mechanism and artificial sense check mechanism also to differ greatly, truly do not possess mankind's similar " perceptional function ", and from the paper periodical delivered and the patent of having authorized, can know and have people that multi-information merging technology is applied to food, the Non-Destructive Testing aspect of the qualities such as agricultural product, as " the quality of famous tea intelligent equipment based on multi-sensor data fusion evaluates method " (ZL200910232916.9) and " a kind of food sensory quality assessment method and system " (ZL201110142990.9), but on multiple sensor information fusion, can find it is substantially all simply connected by several sensor technology, the heat transfer agent larger to physical significance span carries out arbitrary decision superposition, this is not fusion truly in fact, also differ greatly with the perception behavior of the mankind.Human perception behavior is a very complicated process, the topological structure formed by sensory neuron, nervous centralis and cerebral cortex sensory area.Current research shows that sensory neuron end is connected with a kind of star-shaped glial cell, causes the spatial joins of this topological structure very complicated, and the perception information that different sense organ produces exists intersection mutually in transmitting procedure; That is, in sensory analysis, the multiple organ acting in conjunction of the mankind, to influence each other.
Given this, the invention provides a kind of sensory method responding to fusion across perception multi information alternately realizing simulating people.
Summary of the invention
Compared with single detection means, the application of Multi-information acquisition in Intelligent Food organoleptic examination, the information that detection can be made to obtain more comprehensively, result is more objective, and more press close to the perception of the mankind.How to realize the utilizing technology of Multi-information acquisition to carry out food quality intelligent Evaluation and become patent one guardian technique place of the present invention.
The technical scheme that this patent provides is: respond to fusion method alternately across perception information in the bionical evaluation of a kind of Intelligent Food, utilize dissimilar biomimetic sensor from sense of smell, vision, the nose of sense of taste aspect simulation people, eye, tongue perceptual organ, extract the characteristic variable of food, obtain a matrix, comprise the heat transfer agent of different physical significance and different dimension, magnitude, by principal component analysis (PCA), all characteristic variables extracted are recombinated, obtain the major component PC that several are mutually orthogonal
i, the mutual induction realizing dissimilar heat transfer agent is merged; Utilize front n major component PC1, PC2, PC3 ... the sense of smell that PCn checks artificial sense, vision and sense of taste score return, and obtain the dummy variable of corresponding sense of smell, vision and the sense of taste respectively; Finally using the input of these dummy variables as decision system, build non-linear decision system by Genetic Neural Network Method, achieve sense of smell, vision, the sense of taste across detecting sensor information interaction induction fusion.
The concrete steps of the inventive method are:
(1) olfactory information collection, selects the nose of olfactory sensor equipment simulating people to gather the odiferous information of food, obtains p the characteristic variable a that can reflect food smell information according to different olfactory sensor array
1, a
2, a
3... .a
p;
(2) visual information collection, the eyes selecting vision sensor to simulate people gather the exterior quality information of food, carry out color characteristic and analysis of texture, obtain q the characteristic variable b that can reflect appearance information to outward appearance quality information
1, b
2, b
3... b
q; (3) collection of sense of taste information, selects the tongue of taste sensor equipment simulating people to gather the flavour information of food, obtains t the characteristic variable c that can reflect food taste
1, c
2, c
3... c
t;
(4) characteristic variable that (1)-(3) obtain is combined into the matrix that m is capable, (p+q+t) arranges;
(5) adopt the method for principal component analysis (PCA) to complete purification to above-mentioned characteristic variable, dimensionality reduction and screening process, principal component analysis (PCA) is recombinated to former all characteristic variables in higher-dimension Virtual Space, obtains the major component that several are mutually orthogonal;
(6) front n major component PC1, PC2, PC3 is extracted ... PCn, by multiple linear regression set up sense of smell, vision, the sense of taste three types score dummy variable L1, L2, L3 respectively to a front n major component PC1, PC2, PC3 ... the equation of linear regression of PCn;
(7) score dummy variable L1, L2, L3 of the sense of smell of all samples, vision, the sense of taste three types and the artificial sense that corresponds are evaluated scoring rank and substitute into multi-sensor data fusion model based on BP-ANN as input layer, sample is divided into calibration set and forecast set two groups at random according to the ratio of 3:2;
(8) according to BP-ANN Fusion Model Output rusults, check the height of forecast set discrimination, verify the stability of institute's established model.
Adopt in this patent and set up dummy variable: (1) olfactory information collection, select the nose of olfactory sensor equipment simulating people to gather the odiferous information of food, obtain the characteristic variable a that can reflect food smell information according to different olfactory sensor array
1, a
2, a
3... .a
p.(2) visual information collection, selects the eyes of vision sensor technical modelling people to gather the exterior quality information of food, carries out color characteristic and analysis of texture, obtain the characteristic variable b that can reflect appearance information to outward appearance quality information
1, b
2, b
3... b
q.(3) collection of sense of taste information, selects the tongue of taste sensor equipment simulating people to gather the flavour information of food, obtains the characteristic variable c that can reflect food taste
1, c
2, c
3... c
t.The heat transfer agent of obtain three kinds of different physical significances and different dimension, magnitude is combined into the huge matrix that m is capable, (p+q+t) arranges:
In the face of so huge data, be usually mapped to a higher dimensional space according to unified approach, and take out the fusion process that several Virtual vector have come across perception information.This patent adopts the method for principal component analysis (PCA) (PCA) to come purification to characteristic variable, dimensionality reduction and screening process, and principal component analysis (PCA) is recombinated to former all characteristic variables in higher-dimension Virtual Space, obtains the major component that several are mutually orthogonal.Patent of the present invention selects a front n major component PC1, PC2, PC3 with the height of variance contribution ratio ... PCn, the quantity of information that can represent food quality more than 90 percent with it is for the best.Obviously, after principal component analysis (PCA) process, can obtain:
PC1=k
11a
1+k
12a
2+k
13a
3……+k
1pa
p+k
1(p+1)b
1+k
1(p+2)b
2+……k
1(p+q)b
q+k
1(p+q+1)c
1+k
1(p+q+2)c
2+……k
1(p+q+t)c
t,
PC2=k
21a
1+k
22a
2+k
23a
3……+k
2pa
p+k
2(p+1)b
1+k
2(p+2)b
2+……k
2(p+q)b
q+k
2(p+q+1)c
1+k
2(p+q+2)c
2+……k
2(p+q+t)c
t,
PC3=k
31a
1+k
32a
2+k
33a
3……+k
3pa
p+k
3(p+1)b
1+k
3(p+2)b
2+……k
3(p+q)b
q+k
3(p+q+1)c
1+k
3(p+q+2)c
2+……k
3(p+q+t)c
t,
PC4=k
41a
1+k
42a
2+k
43a
3……+k
4pa
p+k
4(p+1)b
1+k
4(p+2)b
2+……k
4(p+q)b
q+k
4(p+q+1)c
1+k
4(p+q+2)c
2+……k
4(p+q+t)c
t,
………………………………………………………………………………………………………………………
………………………………………………………………………………………
PCn=k
n1a
1+k
n2a
2+k
n3a
3……+k
npa
p+k
n(p+1)b
1+k
n(p+2)b
2+……k
n(p+q)b
q+k
n(p+q+1)c
1+k
n(p+q+2)c
2+……k
n(p+q+t)c
t
Each major component is formed by all transduction feature variablees in theory, embodies interactivity.Then front n major component PC1, PC2, PC3 is extracted ... PCn, by multiple linear regression (MLR) set up sense of smell, vision, the sense of taste three types score dummy variable L1, L2, L3 respectively to a front n major component PC1, PC2, PC3 ... the equation of linear regression of PCn:
Virtual sense of smell variables L 1=a+b
11pC1+b
12pC2+b
13pC3+b
14pC4+ ... + b
1npCn;
Virtual vision variables L 2=a+b
21pC1+b
22pC2+b
23pC3+b
24pC4+ ... + b
2npCn;
Virtual sense of taste variables L 3=a+b
31pC1+b
32pC2+b
33pC3+b
34pC4+ ... + b
3npCn
The sense of smell obtained by structure, vision, the sense of taste three score dummy variables, the numeral just truly achieving mutual induction between human olfactory, vision, the sense of taste is bionical.
Because the corresponding relation between the characteristic variable of detecting sensor and sensory quality of food is not simple linear relationship, usually non-linear to genetic neural network, fuzzy support vector machine etc. means are incorporated in Fusion Model foundation, the Fusion Model obtained is close with human perception behavior as much as possible.This patent adopts error backward propagation method (BP-ANN) means to be dissolved in model foundation and goes.Using the true sensory evaluation score (generally providing evaluation score by specialized review expert) of three dummy variables and people as input layer, be updated in BP-ANN model and go, obtain food quality based on the comprehensive distinguishing result across detecting sensor information fusion technology, so as the organs such as the eye of human simulation, nose, tongue to food quality as comprehensive sensory evaluation.
The present invention utilizes modern data method for digging to purify and dimensionality reduction raw data, from the incoherent mutually three class sensor information data of magnanimity, screen characteristic variable; Itself is coupled without practical significance characteristic variable and corresponding artificial sense result by design intelligent learning algorithm, gives its corresponding perceptional function, resolves with realization character variable perception meaning.This patent provides and the heat transfer agent of different physical significance and different dimension, magnitude is mapped to a higher dimensional space according to unified approach, and takes out several score Virtual vector and come to merge across perception information.In view of the corresponding relation between the characteristic variable of detecting sensor and sensory quality of food is not often simple linear relationship, research is planned the non-linear means such as genetic neural network, fuzzy support vector machine and is incorporated in Fusion Model foundation, and the Fusion Model obtained is close with human perception behavior as much as possible.
Embodiment
In the present embodiment, this patent is intended adopting many biomimetic sensors integration technology to be applied to the intellectualized detection of tealeaves millet paste quality, the large sense organ of the eye of simulation people, nose, tongue three.Flow process, as Fig. 1, gathers the transduction feature information of different physical significance and different dimension, magnitude by the sensor of three types.Multi-sensor information gatherer process: (1) olfactory information collection, the nose selecting Electronic Nose instrument and equipment (the PEN3 Electronic Nose of German AIRSENSE company) to simulate people gathers the odiferous information of millet paste, the information that can reflect millet paste smell is obtained according to different olfactory sensor array, extract 3 eigenwerts such as each sensor signal maximal value, minimum value and average, electric nasus system has 10 mos sensors, can extract so altogether 30 odor characteristics variablees; (2) visual information collection, colour examining colour-difference-metre (the full-automatic colour examining colour-difference-metre of DC-P3 type) is selected to gather the color and luster information of millet paste, signature analysis is carried out to millet paste color and luster information, obtain reflecting the aberration brightness L of soup look information, red green degree a, Huang Lan degree b, △ L, △ a, △ b and △ E value, and calculate its derivative value, comprise form and aspect b/a, tone chroma Cab, color saturation Sab and hue angle Hab, totally 11 color and luster characteristic variables; (3) collection of sense of taste information, the tongue selecting electronic tongue instrument equipment (ASTREE II electronic tongues of French Alpha MOS company) to simulate people gathers the flavour information of millet paste, obtain the 7 working electrode signals that can reflect millet paste flavour information, extract sensor stabilization value, then a sample obtains 7 taste characteristics variablees.
Utilize the raw data of principal component analysis (PCA) (PCA) to obtained three kinds of sensor informations to purify and dimensionality reduction, from mass data, screen characteristic variable.The multi information of different physical significance and different dimension, magnitude is mapped in a higher dimensional space according to unified approach to go by PCA, realize the mutual restructuring between multiple mutual incoherent characteristic variable information, obtain the major component that several are mutually orthogonal, vision, sense of smell, the sense of taste three score dummy variables are set up respectively to the equation of linear regression of front 6 major components of PC1, PC2, PC3, PC4, PC5, PC6 again, from truly accomplishing that the mutual induction of many heat transfer agents is merged by multiple linear regression (MLR).Design intelligent learning algorithm, adopts some nonlinear pattern recognition algorithms itself to be coupled without practical significance characteristic variable and corresponding artificial sense result, gives its corresponding perceptional function, resolve with realization character variable perception meaning.Artificial sense evaluation score is generally made up of the evaluation expert group of specialty, comprehensively provide evaluation score according to millet paste soup look, fragrance, flavour according to artificial sense test stone (as table 1), then according to integrated level partitioning standards, millet paste is divided into 1 grade, 2 grades, 3 grades three different quality grades (as shown in table 2).
Table 1 artificial sense test stone
PTS=0.4*X1+0.3*X2+0.3*X3.
Table 2 integrated level partitioning standards
Integrated level |
PTS scope |
1 grade |
90-99 |
2 grades |
80-89 |
3 grades |
70-79 |
This patent adopts error backward propagation method (BP-ANN) means to be dissolved in model foundation and goes.The true perception of three score dummy variables and people is evaluated scoring rank as input layer, be updated in BP-ANN model and go, the Fusion Model obtained is close with human perception behavior as much as possible, so as the organs such as the eye of human simulation, nose, tongue to millet paste quality as comprehensive sensory evaluation.Realize making reasonable, standard compliant evaluation result across bionical sensing technology to millet paste.
Multisensor Multi-information acquisition intelligent Evaluation result: select 45 sample (Three Estates, each grade 15 samples) substitute into BP-ANN Modling model as calibration set, choose 30 sample (Three Estates, each grade 10 samples) stability of institute's established model is verified as forecast set, obtained predict the outcome for, the goodness of fit evaluating level results with artificial sense is 90%.