CN102564993A - Method for identifying rice varieties by using Fourier transform infrared spectrum and application of method - Google Patents
Method for identifying rice varieties by using Fourier transform infrared spectrum and application of method Download PDFInfo
- Publication number
- CN102564993A CN102564993A CN2011104570886A CN201110457088A CN102564993A CN 102564993 A CN102564993 A CN 102564993A CN 2011104570886 A CN2011104570886 A CN 2011104570886A CN 201110457088 A CN201110457088 A CN 201110457088A CN 102564993 A CN102564993 A CN 102564993A
- Authority
- CN
- China
- Prior art keywords
- rice
- model
- samples
- sample
- fourier transform
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates a method for identifying rice varieties by using a Fourier transform infrared spectrum and application of the method. The method comprises the following steps of: determining Fourier transform infrared spectrum determining parameters of rice samples; establishing an identification model of the rice varieties; and carrying out unknown sample identification and the like. The method disclosed by the invention can be used for identifying rice very rapidly and accurately, thereby being good for optimization of a brewing process, control of production flows and stabilization of yellow rice wine quality.
Description
[technical field]
The invention belongs to agricultural technology field.More specifically, the present invention relates to utilize the method for Fourier transform infrared spectroscopy identification rice variety, also relate to the recognition methods in the glutinous rice place of production.
[background technology]
Rice is the main material of yellow rice wine brewage, and the rice in the different cultivars and the place of production exists than big-difference at aspects such as chemical composition, physical characteristics and microbial populations, has influence on the quality of yellow rice wine brewage then.Rice to the different cultivars and the place of production is discerned fast and accurately, helps optimizing brewage process, and the yellow rice wine quality is stablized in the control production procedure.
CN200810112084 discloses the automatically method of identification of a kind of yellow rice kernel, and this comprises rice appearance is placed image acquisition device, gathers and read original image information, separating background and rice appearance; Read original chrominance information R, G, the B value of every rice,, R, G, B chromatic value are converted into even color model chromatic value L through desired color model XYZ
*A
*B
*Select to judge characteristic chromatic value B
*, if characteristic chromatic value B in all pixels of every rice
*Surpass the colourity threshold value, and proportion judges that this rice is a yellow rice kernel during greater than area threshold.Wherein colourity threshold value and area threshold are set up on their own by the kind and the place of production of operator according to rice.CN200810111705 discloses the method that a kind of rice granule type detects, and this method comprises rice appearance is placed image acquisition device, gathers and read original image information, separating background and rice appearance; The whole grain of rice of identification and broken rice; The whole grain of rice quantity that rice variety during system is provided with according to parameter is required with detection is selected requirement, point by point search at random in the whole grain of rice.
At present; Fourier transform infrared spectroscopy (Fourier Transform Infrared Spectroscopy; FTIR) be the vibration and the rotation spectrum of representative functional group in a kind of main research material molecule, the infrared spectrum sample preparation is simple fast, good reproducibility, sample is not damaged, contains much information.Utilize Fourier transform infrared spectroscopy diffuse reflection method (FTIR-DR) to gather spectrogram; And combination chemometrics method; Employing stays a cross verification to set up principal component analysis (PCA) (PCA) model, is that index is discerned the rice in the different cultivars and the place of production with discrimination and reject rate.
This method has been confirmed grain size, sample ratio, sampling parameter and the characteristic spectrum scope after optimum rice sample is pulverized, and through the optimization pre-service of spectrogram, adopts chemometrics method to set up the model of cognition in the rice variety and the place of production.
[summary of the invention]
[technical matters that will solve]
The purpose of this invention is to provide a kind of method of utilizing Fourier transform infrared spectroscopy identification rice variety.
Another object of the present invention provides the purposes of said identification rice variety method.
[technical scheme]
The present invention realizes through following technical proposals.
The present invention relates to a kind of method of utilizing Fourier transform infrared spectroscopy identification rice variety.
This method comprises the steps:
A) confirm the location parameter of rice sample Fourier transform infrared spectroscopy;
B) set up the model of cognition of rice variety
300 rice samples are divided into two collection at random, and wherein 250 rice samples are used for the foundation of model as training set, and 50 rice samples are as inspection set;
(1) the selection analysis of characteristic wavelength
These rice samples carry out Fourier transform infrared spectroscopy and measure; The infrared spectrogram of gathering is imported Unscrambler software; Handle through Matrix, obtain the infrared spectrum three-dimensional matrice figure of these rice samples, select the proper vector of its characteristic wavelength as pattern-recognition;
(2) pre-service of data
These proper vectors are carried out level and smooth, baseline correction and the standard normal variable conversion pre-service automatically of spectrum Savitzky-Golay, carry out PCA again and analyze and set up model of cognition;
(3) foundation of rice variety model of cognition
Adopt the NIPALS computing method and stay a proof method that valid data in the training set are carried out the PCA analysis alternately, obtain the three-dimensional shot chart of PCA of these rice, set up the SIMCA model of cognition of every kind of rice on this basis;
C) unknown sample identification
Adopt the pattern-recognition projecting method of SIMCA that unknown sample is predicted; Investigate the distance relation between contribution rate in the mould, intermode ability to see things in their true light, model spacing model parameter and unknown appearance and model in the projection process; Finally draw its recognition result, its discrimination and reject rate are represented:
A preferred embodiment of the invention; Use FTIS; Through the ir data of gathering; Analyze the influence of dilution ratio, scanning times and the resolution of rice sample powder size, rice and dilution matrix, confirm to measure the basic parameter of rice infrared spectrum the rice infrared spectrum.
According to another kind of preferred implementation of the present invention, described dilution matrix is potassium bromide.
According to another kind of preferred implementation of the present invention, the dilution ratio 1: 9 of rice sample powder size 200 orders, rice and dilution matrix, scanning times 16 times and resolution 4cm
-1
According to another kind of preferred implementation of the present invention, the operating conditions of FTIS is following: minimum resolution 0.5cm
-1, EverG10TM Mid-Infrared Source light source, Smart Diffuse Reflection annex, DTGS (deuterium thiosulfuric acid three glycosides peptides)/KBr detecting device.
According to another kind of preferred implementation of the present invention, described rice is long-grained nonglutinous rice, polished rice or glutinous rice.
According to the present invention, the above-mentioned method of utilizing Fourier transform infrared spectroscopy to discern rice variety is used for the identification in the glutinous rice place of production.
A preferred embodiment of the invention when setting up glutinous rice place of production model of cognition, is gathered the infrared spectrum of 200 glutinous rice samples, at random with 150 samples as training set, 30 samples are as forecast set.
Below the present invention will be described in more detail.
The present invention relates to a kind of method of utilizing Fourier transform infrared spectroscopy identification rice variety.
This method comprises the steps:
A) confirm the location parameter of rice sample Fourier transform infrared spectroscopy.
FTIS is based on the infrared light after interfering is carried out the principle of Fourier transform and a kind of infrared spectrometer of developing, and it mainly is made up of infrared light supply, diaphragm, interferometer (beam splitter, index glass, horizontal glass), sample chamber, detecting device and various ir reflector, laser instrument, control circuit board and power supply.Can carry out qualitative and quantitative analysis to sample, be widely used in technical fields such as medicine, chemical industry, environmental protection.The principal feature of FTIS is that this instrument does not have grating or prismatic decomposition device, has reduced optical loss, and further increases light signal through interfering, and the radiation intensity that therefore arrives detecting device is big, and signal to noise ratio (S/N ratio) is high.The signal of the Fourier pair light that FTIS adopts is handled, and the error of having brought when having avoided the motor-driven grating beam splitting is so reappearance is relatively good.FTIS carries out data acquisition according to all band; The spectrum that obtains is to the result after repeatedly data acquisition is asked on average; And the data acquisition of accomplishing one whole only needs one to the several seconds; The very narrow frequency range of color dispersion-type instrument mensuration in a flash then in office, the one whole data acquisition needs 10-20 minute.
In the present invention; FTIS is a product sold in the market, for example the FTIR-650 FTIS produced of Gangdong, Tianjin, the WQF-510 FTIS that Beijing Rayleigh is produced, the Nicolet6700 FTIS that U.S. Thermo Fisher produces; Tensor 37 FTISs that Germany Bruker Optics produces.
The rice that uses in test is 300 rice samples from ten different regions such as Lianyun Harbour, Jiangsu Province, Daqing, Heilungkiang, Anqing, Anhui Province, Hubei Province's Xiaogan City, Hubei Province's Wuxue City, Hangzhou, Zhejiang province city, Wuhu City, Anhui Province, Bengbu, Anhui Province, Shaanxi Province's Hanzhong City, Sichuan Province's Mianyang Cities.These samples are divided into polished rice (J), glutinous rice (N), long-grained nonglutinous rice (X) by kind, and glutinous rice is numbered HX, ZH, HW, SH, AW, AB according to the place of production, and specifying information is listed in table 1.
Table 1 different cultivars and place of production rice sample and numbering
Annotate: J, N, X are that variety classes is represented in the capitalization of getting the first phonetic alphabet of polished rice, glutinous rice, long-grained nonglutinous rice respectively in the table, the numbering of the different rice of letter back numeral; Place of production numbering is made up of the initial of place, place of production province and regional lead-in.
The present invention uses FTIS; Through the ir data of gathering; Analyze the influence of dilution ratio, scanning times and the resolution of rice sample powder size, rice and dilution matrix, confirm to measure the basic parameter of rice infrared spectrum the rice infrared spectrum.
Described dilution matrix is potassium bromide.
The operating conditions of FTIS is following: minimum resolution 0.5cm
-1, EverG10TM Mid-Infrared Source light source, Smart Diffuse Reflection annex, DTGS (deuterium thiosulfuric acid three glycosides peptides)/KBr detecting device.
Data analysis is based on the software platform of Unscrambler 9.7.
Particularly, FTIR-DR diffuse reflection method is applicable to the infrared spectrum collection of powdered sample, but sample particle is big or small, uniformity coefficient, the sample ratio during sampling, and sampling parameter all can influence the quality of infrared spectrum and the reappearance of infrared spectrum.For avoiding introducing error in sample preparation and sampling element, have influence on subsequent builds model accuracy and accuracy, optimize the sample preparation parameter and be very important.The inventor studies the influence of rice infrared spectrum with regard to factors such as the dilution ratio of rice sample powder size, rice and dilution matrix, scanning times, resolution.
1, different rice powder sizes are to the influence of infrared spectrum.
Pulverize rice sample; Use 60 orders, 100 orders, 200 mesh sieves to sieve and obtain three kinds of varigrained rice powdered samples; These samples were preserved 24 hours under constant temperature and humidity; The rice sample that obtains mixes sample preparation according to this rice sample with KBr mass ratio 10%, and scanning times is 16 times under above-mentioned FTIS operating conditions, resolution 4cm
-1, obtain the diffuse reflectance IR figure of accompanying drawing 1 like this.
Can be found out that by accompanying drawing 1 the order number of rice powder is high more, the signal to noise ratio (S/N ratio) of infrared spectrum is high more, and spectral line is level and smooth more.When the rice powder size reached 200 orders, its spectrogram noise obviously reduced, at rice diffuse reflectance infrared spectroscopy district 800cm
-1～1800cm
-1Wave band, its noise is eliminated basically.If further improve the rice sample granularity, the pulverizing difficulty adds beats, and its energy consumption also increases substantially; The particle relative surface area increases, and has also increased the hydroscopicity of sample, makes the sample signal distortion; Influence the quality of infrared spectrum, so rice sample powder size 200 orders preferably.
2, in the IR spectroscopy sample sample concentration to the influence of infrared spectrum
According to the present invention, described IR spectroscopy sample should be appreciated that it is to carry out the employed sample of infrared spectrum measurement, and it is substrate composed with dilution by material rice powder to be determined.
Described sample concentration should be appreciated that it is the content of test substance rice powder in dilution matrix in the employed sample of infrared spectrum measurement.
The IR spectroscopy sample concentration also is one of key factor that influences the infrared spectrum plot quality.Sample concentration is too big, direct reflection can occur, the loss luminous energy, and high concentration also can make bands of a spectrum broaden, and resolution descends, even the hypersorption phenomenon occurs, can not give full expression to out the information of chemical substance composition and content in the rice; Sample concentration is low excessively, can cause absorbing not obvious, and signal is not enough, can't express the difference of different sample rooms effectively.
The present invention uses in the IR spectroscopy normally used KBr as dilution matrix; Its absorption spectrum is measured the background of spectrum as it; 200 order rice powdered samples and KBr powder are according to 20%, 10%, 5% mass percent mixing; Be pressed into tablet, under above-mentioned FTIS operating conditions, measure, obtain the infrared spectrogram shown in accompanying drawing 2.
Can find out by accompanying drawing 2; The mass percent of 200 order rice powdered samples and KBr powder is that 10% o'clock gained collection of illustrative plates signal to noise ratio (S/N ratio) is best, and light absorption value meets the range of linearity of Beer-Lambert law in 0.3～0.8 scope; And the mass percent of 200 order rice powdered samples and KBr powder be 5% o'clock infrared spectroscopy signals very a little less than; 20% o'clock sample too high levels, infrared spectroscopy signals is strong excessively, all is unfavorable for infrared spectrum measurement; Therefore, the mass percent of 200 order rice powdered samples and KBr powder preferably 10%.
3, scanning times is to the influence of infrared spectrum
According to the present invention, scanning times should be appreciated that it is infrared spectrometer carries out replication to an IR spectroscopy sample under above-mentioned FTIS operating conditions number of times, and it is an important parameters of spectra collection.Under the constant situation of other condition determination, there is a kind of positively related relation in the signal to noise ratio (S/N ratio) of infrared spectrometer scanning times and infrared spectrogram, and promptly scanning times is many more, and the signal to noise ratio (S/N ratio) of spectrum is high more.But in actual infrared analysis, increase scanning times and just increase minute, can also increase its cost of determination because of the infrared spectrum measurement environment change can increase error at measurment simultaneously.Therefore, normally adopt the least possible IR spectrum scanning number of times guaranteeing that infrared spectrogram has under the prerequisite of good signal to noise ratio (S/N ratio).Scanning times is set is respectively 8,16,32 o'clock resultant infrared spectrograms shown in accompanying drawing 3.
Can find out that by accompanying drawing 3 when an IR spectroscopy sample was carried out multiple scanning 16 times, the signal to noise ratio (S/N ratio) of its infrared spectrogram was best.
4, resolution is to the influence of infrared spectrum
According to the present invention, described resolution should be appreciated that it is the minimum wavelength interval that when the infrared spectrometer detecting device receives infrared spectroscopy signals, can differentiate, and it directly influences the fine structure of spectrogram.Resolution is high more, and spectral composition is meticulous more, and minute is long more, and the data volume of infrared spectrum is big more; Resolution is low, and then the sampling time shortens, and the data volume of infrared spectrum reduces, and the infrared spectrum structure is not meticulous.In the present invention, be set at 2cm respectively in resolution
-1, 4cm
-1, 8cm
-1Condition under carried out infrared spectrum measurement.Its rice infrared spectrogram is shown in accompanying drawing 4.
Can find out by accompanying drawing 4, when resolution is 4cm
-1The time infrared spectrogram signal to noise ratio (S/N ratio) reached the analysis requirement, therefore confirm that resolution is 4cm
-1
Preferably, the basic parameter of mensuration rice infrared spectrum is dilution ratio 1:9, scanning times 16 times and the resolution 4cm of rice sample powder size 200 orders, rice and dilution matrix
-1
Described rice is long-grained nonglutinous rice, polished rice or glutinous rice.
B) set up the model of cognition of rice variety
Pattern-recognition; Be called pattern classification again; Handle and analyze characterizing things or phenomenon various forms of (numerical value, literal with logical relation) information; With the process that things or phenomenon are described, recognize, classified and explain, be a kind of sample to be carried out the process of cluster, be the comprehensive utilization of mathematics, information science and computer science.(Soft Independent Modeling of Class Analogy SIMCA) is one of mode identification method of using always to soft independent modeling classification.
SIMCA is a kind of mode identification method that supervision is arranged; Earlier each type known sample in the training set is carried out PCA respectively and analyze and set up mathematical model; Then unknown sample and the model of having set up are carried out match, confirm that which kind of unknown sample belongs to or do not belong to any one type.SIMCA comprises two key steps: the PCA regression model of setting up each type sample; Utilize model that unknown sample is carried out match, confirm its classification.
Principal component analysis (PCA) (PCA) is the core of SIMCA identification; PCA can carry out projection to the complicated mutual relationship of multidimensional space data to lower dimensional space and obtain simplifying, and puts in order Useful Information among several major components and rejects garbages such as noise, error through dimensionality reduction, in addition; Mutually orthogonal between the major component; Can overcome the information overlap that former variable multiple correlation causes, help to extract to greatest extent Useful Information, set up outstanding mathematical model.
When data analysis, PCA is with the spectrum matrix A
M * pResolve into the T in the formula (1)
M * fAnd P
F * pTwo than the apposition of minor matrix and residual matrix E with:
A
m×p＝T
m×fP
f×p+E (1)
A in the formula
M * pBe spectrum matrix, T
M * fBe score (Score) matrix, P
F * pBe load (Loading) matrix, E is regarded as " residual error ", dimension and A
M * pIdentical, m is the sample number, and p is the spectral variables number, and f is the major component number.Score promptly is major component matrix (new variables) in PCA analyzes, and load then is the correlation matrix of major component and spectrum matrix, and the apposition of two matrixes just in time obtains the dimension identical with the spectrum matrix, adds the matrix E that is used as residual error and just can reduce the spectrum matrix.What adopt when PCA analyzes in addition is nonlinear iteration offset minimum binary algorithm (NIPALS).
300 rice samples are divided into two collection at random, and wherein 250 rice samples are used for the foundation of model as training set, and 50 rice samples are as inspection set.The data of different research objects have different characteristics, need carry out necessary processing to original spectroscopic data for obtaining good discrimination model and prediction effect, like the selection of characteristic wavelength, data pre-service etc.
(1) the selection analysis of characteristic wavelength
Rice sample carries out Fourier transform infrared spectroscopy and measures.Different cultivars rice sample and numbering are listed in the table 2.
Table 2 different cultivars rice sample and numbering
The infrared spectrogram of gathering is imported Unscrambler software, handle, obtain the infrared spectrum three-dimensional matrice figure of these rice samples, select the proper vector of its characteristic wavelength as pattern-recognition through Matrix.
Described Unscrambler software (multivariate data analysis software) is a complete multivariate data analysis and experimental design software; It comprises powerful statistical method; Like PCA, multivariate curve is differentiated instrument (MCR), PLS recurrence, 3-Way PLS recurrence, cluster (K-Means), SIMCA and PLADA classification etc.
Infrared spectrum three-dimensional matrice figure lists in the accompanying drawing 5.Can find out that by accompanying drawing 5 infrared spectrogram of these three kinds of rice samples is at 1190cm
-1～1465cm
-1(A district) and 2846cm
-1～2936cm
-1(B district) two wave bands have stronger absorption, and absorbance difference is also bigger, and these two wave bands mainly are the absorption regions of components such as carbohydrate, protein (amino acid), fat, other nitrogenate, therefore elect it proper vector of pattern-recognition as.
(2) pre-service of data
These proper vectors are carried out the proper vector pre-service listed like table 3, i.e. level and smooth, the baseline correction and standard normal variable conversion pre-service automatically of spectrum Savitzky-Golay carried out the PCA analysis again and set up model of cognition.
Table 3: the pre-service of proper vector
(3) foundation of rice variety model of cognition
Adopt the NIPALS computing method and stay a proof method that valid data in the training set are carried out the PCA analysis alternately, obtain the three-dimensional shot chart of PCA of these rice, set up the SIMCA model of cognition of every kind of rice on this basis.
The three-dimensional shot chart of PCA is shown in accompanying drawing 6.From accompanying drawing 6, can see the effect of three kinds of rice good discrimination intuitively; Relatively independent and the comparatively tangible aggregation zone of each self-forming of long-grained nonglutinous rice, polished rice and glutinous rice; Reached the cluster requirement of it being carried out the SIMCA modeling, also definite sample preparation parameter and the original spectrum of explanation handled the effect of all having brought into play.
Respectively long-grained nonglutinous rice, polished rice, glutinous rice in the training set are carried out PCA and analyze, obtain optimum modeling number of principal components and confirm as 4,4,6 respectively, set up the SIMCA model of cognition of every kind of rice on this basis.
C) unknown sample identification
Adopt the pattern-recognition projecting method of SIMCA that unknown sample is predicted, investigate the distance relation between contribution rate in the mould, intermode ability to see things in their true light, model spacing model parameter and unknown appearance and model in the projection process, finally draw its recognition result.The SIMCA sciagraphy obtains predicting the outcome of unknown sample, representes with discrimination and reject rate, and discrimination and reject rate are the confidence levels of differentiating unknown sample between the reflection class model.
Described discrimination should be appreciated that being investigated unknown sample drops on the ratio in the correct class model zone; Represented suc as formula (1); And reject rate should be appreciated that being investigated class model does not belong to the refusal degree of such unknown sample to other; The sample that does not promptly belong to such drops on the outer probability of such model area, and is represented suc as formula (2).Discrimination and reject rate numerical value are more near 100%, and the prediction effect of class model is good more.
Discrimination and reject rate are represented as follows:
Table 4 adopts discrimination and the reject rate of SIMCA projecting method to whole 50 unknown rice samples (5 samples of every kind of rice) for the present invention, selects the level of signifiance of α=5% for use.
Table 4 unknown sample predicts the outcome
Find out that from table 4 every kind of rice is all belonged under the kind separately on 100% ground, and model 100% ground " refusal " of every kind of rice the rice of other kinds, explain that the recognition effect of recognition mode of the rice variety of being set up is remarkable.
According to the present invention, the above-mentioned method of utilizing Fourier transform infrared spectroscopy to discern rice variety can be used for the identification in the glutinous rice place of production.
Adopt the above-mentioned method of Fourier transform infrared spectroscopy identification rice variety of utilizing to carry out the identification in the glutinous rice place of production, its step is following:
1, sets up the model of cognition in the glutinous rice place of production
Gather the infrared spectrum of 200 glutinous rice samples, at random with 150 samples as training set, 30 samples are as forecast set.These glutinous rice samples carry out Fourier transform infrared spectroscopy and measure; The infrared spectrogram of gathering is imported Unscrambler software; Handle through Matrix, obtain the infrared spectrum three-dimensional matrice figure of these glutinous rice samples, select the proper vector of its characteristic wavelength as pattern-recognition.Choose 1751-1685cm
-1, 1375-1180cm
-1The interval data of wave band are carried out modeling.According to like foregoing method, Fourier transform infrared spectroscopy figure is carried out the data pre-service.
2, the foundation of glutinous rice place of production model of cognition
According to like foregoing method, adopt the NIPALS computing method and stay a proof method that useful data in the training set is carried out the PCA analysis alternately, obtain the three-dimensional shot chart of PCA of six kinds of place of production glutinous rice, its result is as shown in Figure 7.The separate distribution of various glutinous rice shows each other and can well separate, can set up good model of cognition.
The modeling number of principal components of the optimum in each place of production (AB, AW, HW, HX, SH, ZH) confirms as 2,5,4,3,4,4 respectively, sets up the SIMCA model of cognition of different places of production glutinous rice on this basis.
3, prediction unknown sample
Forecast set sample substitution model is carried out match, obtain the discrimination and the reject rate of different places of production glutinous rice, its result is as shown in table 6.Unknown place of production glutinous rice sample predicts the outcome under the level of significance of α=5%.Glutinous rice discrimination=100% in each place of production, reject rate >=88% explains that institute's established model can discern the glutinous rice place of production well.
Table 6 unknown sample predicts the outcome
[beneficial effect]
The invention has the beneficial effects as follows: the present invention utilizes the Fourier transform infrared spectroscopy technology; Through the ir data of gathering; Analyze the influence of dilution ratio, scanning times and the resolution of rice sample powder size, rice and dilution matrix to the rice infrared spectrum; Confirm to measure the basic parameter of rice infrared spectrum,, adopt chemometrics method to set up the model of cognition in the rice variety and the place of production through the optimization pre-service of spectrogram.Method of the present invention can be discerned rice very quickly and accurately, thereby helps optimizing brewage process, and the yellow rice wine quality is stablized in the control production procedure.
[description of drawings]
Fig. 1 is the infrared spectrogram (being 60 orders, 100 orders, 200 orders from top to bottom successively) of the rice under the different grain size;
Fig. 2 is the rice infrared spectrogram (rice content is followed successively by 5%, 10%, 20% from top to bottom) of different sample concentrations;
Fig. 3 is the rice infrared spectrogram (being followed successively by from top to bottom 8 times, 16 times, 32 times) of different scanning number of times;
Fig. 4 is that the spectrogram under different resolution (is followed successively by 2cm from top to bottom
-1, 4cm
-1, 8cm
-1);
Fig. 5 is the infrared spectrum matrix diagram of rice sample;
Fig. 6 is three-dimensional (PCI-PC2-PC3) shot chart of three kinds of rice;
Fig. 7 is three-dimensional (PC1-PC2-PC3) shot chart of different places of production glutinous rice;
Fig. 8 is the infrared spectrum of rice sample.
[embodiment]
Can understand the present invention better through following embodiment.
Embodiment 1: utilize Fourier transform infrared spectroscopy identification glutinous rice kind
This embodiment implementation step is following:
A): measure infrared spectrogram
Use FTIS at minimum resolution 0.5cm
-1, EverG10TMMid-Infrared Source light source, Smart Diffuse Reflection annex, DTGS (deuterium thiosulfuric acid three glycosides peptides)/KBr detecting device operating conditions; Fourier transform infrared spectroscopy sampling parameter according to the table 1-1 that confirms in this instructions lists is measured infrared spectrogram, and its result sees accompanying drawing 8.
Table 1-1: Fourier transform infrared spectroscopy sampling parameter
B): the model of cognition of setting up rice variety
(1) the selection analysis of characteristic wavelength
The infrared spectrogram of gathering is imported Unscrambler software, handle through Matrix, obtain the infrared spectrum three-dimensional matrice figure of these rice samples, figure selects its characteristic wavelength according to this infrared spectrum three-dimensional matrice, and they are 1000-1750cm
-1, 2845-2862cm
-1With 2914-2936cm
-1, with the proper vector of these characteristic wavelengths as pattern-recognition.
(2) pre-service of data
These proper vectors are carried out level and smooth, the automatic baseline correction of spectrum Savitzky-Golay, standard normal variable conversion and first derivation pre-service, carry out PCA again and analyze and set up model of cognition; Preprocess method is listed among the table 1-2.
Table 1-2: the preprocess method of proper vector
(3) foundation of rice variety model of cognition
Adopt the NIPALS computing method and stay a proof method to carry out the PCA analysis alternately; Obtain the three-dimensional shot chart of PCA of these rice; Glutinous rice, polished rice, three kinds of different cultivars rice of long-grained nonglutinous rice are set up the pca model number of principal components and are respectively 4,3,1, set up the SIMCA model of cognition of every kind of rice on this basis.
C): the identification unknown sample
After setting up the SIMCA model glutinous rice is carried out the kind prediction, obtain the result and list among the table 1-3.
Table 1-3: the predicting the outcome of glutinous rice kind
Result by last table clearly illustrates that when the model of cognition that characteristic wavelength of employing this method and pre-service are set up carried out kind identification to unknown sample, the discrimination of glutinous rice was 97%, and reject rate is 90%.
Embodiment 2: combine Fourier transform infrared spectroscopy and SIMCA identification polished rice kind
This embodiment implementation step is following:
A): measure infrared spectrum
Use FTIS at minimum resolution 0.5cm
-1, EverG10TMMid-Infrared Source light source, Smart Diffuse Reflection annex, DTGS (deuterium thiosulfuric acid three glycosides peptides)/KBr detecting device operating conditions; Fourier transform infrared spectroscopy sampling parameter according to the table 2-1 that confirms in this instructions lists is measured infrared spectrogram, and its result sees accompanying drawing 8.
Table 2-1: Fourier transform infrared spectroscopy sampling parameter
B): the model of cognition of setting up rice variety
(1) the selection analysis of characteristic wavelength
The infrared spectrogram of gathering is imported Unscrambler software, handle through Matrix, obtain the infrared spectrum three-dimensional matrice figure of these rice samples, figure selects its characteristic wavelength according to this infrared spectrum three-dimensional matrice, and they are 1190-1465cm
-1, 2845-2862cm
-1With 2914-2936cm
-1,, list among the table 2-2 the proper vector of these characteristic wavelengths as pattern-recognition.
(2) pre-service of data
These proper vectors are carried out the level and smooth and automatic baseline correction pre-service of spectrum Savitzky-Golay, carry out PCA again and analyze and set up model of cognition; Preprocess method is listed among the table 2-2.
Table 2-2: characteristic wavelength and preprocess method
(3) foundation of rice variety model of cognition
Adopt the NIPALS computing method and stay a proof method to carry out the PCA analysis alternately; Obtain the three-dimensional shot chart of PCA of these rice; Glutinous rice, polished rice, three kinds of different cultivars rice of long-grained nonglutinous rice are set up the pca model number of principal components and are respectively 1,1,1, set up the SIMCA model of cognition of every kind of rice on this basis.
C): unknown sample identification
After setting up the SIMCA model polished rice is carried out the kind prediction, obtain the result and list among the table 2-3.
Table 2-3: the predicting the outcome of polished rice kind
Result by last table clearly illustrates that when the model of cognition that characteristic wavelength of employing this method and pre-service are set up carried out kind identification to unknown sample, the discrimination of polished rice was 100%, and reject rate then is 86%.
Embodiment 3: combine Fourier transform infrared spectroscopy and the SIMCA identification glutinous rice place of production
This embodiment implementation step is following:
A): infrared spectrum is measured
Use FTIS at minimum resolution 0.5cm
-1, EverG10TMMid-Infrared Source light source, Smart Diffuse Reflection annex, DTGS (deuterium thiosulfuric acid three glycosides peptides)/KBr detecting device operating conditions; Fourier transform infrared spectroscopy sampling parameter according to the table 3-1 that confirms in this instructions lists is measured infrared spectrogram, and its result sees accompanying drawing 8.
Table 3-1: Fourier transform infrared spectroscopy sampling parameter
B): the model of cognition of setting up the glutinous rice place of production
(1) the selection analysis of characteristic wavelength
The infrared spectrogram of gathering is imported Unscrambler software, handle through Matrix, obtain the infrared spectrum three-dimensional matrice figure of these rice samples, figure selects its characteristic wavelength according to this infrared spectrum three-dimensional matrice, and they are 980-1170cm
-1, 1180-1375cm
-1With 1685-1751cm
-1,, list among the table 3-2 the proper vector of these characteristic wavelengths as pattern-recognition.
(2) pre-service of data
These proper vectors are carried out spectrum Savitzky-Golay9 point is level and smooth, standard normal variable conversion and second order differentiate pre-service, carry out the PCA analysis again and set up model of cognition; Preprocess method is listed among the table 3-2.
Table 3-2: characteristic wavelength and preprocess method
(3) foundation of rice variety model of cognition
Adopt the NIPALS computing method and stay a proof method to carry out the PCA analysis alternately; Obtain the three-dimensional shot chart of PCA of these rice; Different places of production glutinous rice is set up model; The glutinous rice in AB, AW, HW, HX, SH, ZH six places of production is set up the pca model number of principal components and is respectively 2,2,2,2,2,2, sets up the SIMCA model of cognition of different places of production glutinous rice on this basis.
C): the identification unknown sample
After setting up the SIMCA model glutinous rice is carried out the kind prediction, obtain the result and list among the table 3-3.
Table 3-3: the kind of glutinous rice predicts the outcome
Result by last table clearly illustrates that; When the model of cognition that characteristic wavelength of employing this method and pre-service are set up carries out place of production identification to the glutinous rice sample; Except that the discrimination of SH is that all the other all are 100% 80%; Reject rate is then generally on the low side, explains that different wavelengths can obtain different effects with preprocess method.
Claims (8)
1. a method of utilizing Fourier transform infrared spectroscopy identification rice variety is characterized in that this method comprises the steps:
A) confirm the location parameter of rice sample Fourier transform infrared spectroscopy;
B) set up the model of cognition of rice variety
300 rice samples are divided into two collection at random, and wherein 250 rice samples are used for the foundation of model as training set, and 50 rice samples are as inspection set;
(1) the selection analysis of characteristic wavelength
These rice samples carry out Fourier transform infrared spectroscopy and measure; The infrared spectrogram of gathering is imported Unscrambler software; Handle through Matrix, obtain the infrared spectrum three-dimensional matrice figure of these rice samples, select the proper vector of its characteristic wavelength as pattern-recognition;
(2) pre-service of data
These proper vectors are carried out level and smooth, baseline correction and the standard normal variable conversion pre-service automatically of spectrum Savitzky-Golay, carry out PCA again and analyze and set up model of cognition;
(3) foundation of rice variety model of cognition
Adopt the NIPALS computing method and stay a proof method that valid data in the training set are carried out the PCA analysis alternately, obtain the three-dimensional shot chart of PCA of these rice, set up the SIMCA model of cognition of every kind of rice on this basis;
C) unknown sample identification
Adopt the pattern-recognition projecting method of SIMCA that unknown sample is predicted; Investigate the distance relation between contribution rate in the mould, intermode ability to see things in their true light, model spacing model parameter and unknown appearance and model in the projection process; Finally draw its recognition result, its discrimination and reject rate are represented:
2. method according to claim 1; It is characterized in that using FTIS; Through the ir data of gathering; Analyze the influence of dilution ratio, scanning times and the resolution of rice sample powder size, rice and dilution matrix, confirm to measure the basic parameter of rice infrared spectrum the rice infrared spectrum.
3. method according to claim 2 is characterized in that described dilution matrix is potassium bromide.
4. method according to claim 2 is characterized in that dilution ratio 1:9, the scanning times 16 times and resolution 4cm of rice sample powder size 200 orders, rice and dilution matrix
-1
5. method according to claim 2 is characterized in that the operating conditions of FTIS is following: minimum resolution 0.5cm
-1, EverG10TM Mid-Infrared Source light source, Smart Diffuse Reflection annex, DTGS (deuterium thiosulfuric acid three glycosides peptides)/KBr detecting device.
6. method according to claim 2 is characterized in that described rice is long-grained nonglutinous rice, polished rice or glutinous rice.
7. be used for the identification in the glutinous rice place of production according to the described method of utilizing Fourier transform infrared spectroscopy to discern rice variety of each claim among the claim 1-5.
8. method according to claim 7 is characterized in that when setting up glutinous rice place of production model of cognition, gathers the infrared spectrum of 200 glutinous rice samples, at random with 150 samples as training set, 30 samples are as forecast set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110457088.6A CN102564993B (en) | 2011-12-31 | 2011-12-31 | Method for identifying rice varieties by using Fourier transform infrared spectrum and application of method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110457088.6A CN102564993B (en) | 2011-12-31 | 2011-12-31 | Method for identifying rice varieties by using Fourier transform infrared spectrum and application of method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102564993A true CN102564993A (en) | 2012-07-11 |
| CN102564993B CN102564993B (en) | 2015-07-15 |
Family
ID=46410984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110457088.6A Expired - Fee Related CN102564993B (en) | 2011-12-31 | 2011-12-31 | Method for identifying rice varieties by using Fourier transform infrared spectrum and application of method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102564993B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103837492A (en) * | 2014-02-24 | 2014-06-04 | 西北农林科技大学 | Nondestructive detection method for swelled kiwi fruits based on near-infrared spectrum technology |
| CN104198428A (en) * | 2014-08-21 | 2014-12-10 | 中国农业大学 | Method and system for rapidly identifying authenticity of seeds with seed coatings |
| CN104215584A (en) * | 2014-08-29 | 2014-12-17 | 华南理工大学 | Hyper-spectral image technology-based detection method for distinguishing rice growing areas |
| CN105021562A (en) * | 2015-07-06 | 2015-11-04 | 黑龙江八一农垦大学 | Traceable method for rice origin and application thereof |
| CN105675534A (en) * | 2016-03-25 | 2016-06-15 | 北京市农林科学院 | Method for quickly and nondestructively identifying polished grains |
| CN105678632A (en) * | 2016-03-15 | 2016-06-15 | 江西融思科技有限公司 | Method and structure for tracing detonator manufacturer according to detonator tube shell |
| CN105738340A (en) * | 2015-11-05 | 2016-07-06 | 新疆大学 | Quick detection method of lavender essential oil categories on the basis of Fourier-Raman spectrum |
| CN106092822A (en) * | 2016-06-07 | 2016-11-09 | 河南工业大学 | A kind of distinguish waxy with non-waxy rice method |
| CN106198442A (en) * | 2016-09-14 | 2016-12-07 | 南京财经大学 | A kind of spectroscopic analysis methods differentiating Panjin rice |
| CN104215584B (en) * | 2014-08-29 | 2017-01-04 | 华南理工大学 | A kind of detection method distinguishing the rice place of production based on hyper-spectral image technique |
| CN110118748A (en) * | 2019-04-29 | 2019-08-13 | 山西农业大学 | The construction method of the fourier-transform infrared finger-print of different ageing time Shanxi mature vinegars and application |
| CN111401794A (en) * | 2020-04-24 | 2020-07-10 | 江苏傲农生物科技有限公司 | Feed quality control method based on near infrared spectrum |
| CN112161949A (en) * | 2020-09-16 | 2021-01-01 | 贵州国台酒业股份有限公司 | Method for identifying Maotai-flavor liquor brewing process based on infrared spectrum technology |
| CN112485216A (en) * | 2020-11-20 | 2021-03-12 | 华中农业大学 | Multi-source information fusion Thailand jasmine rice adulteration identification method |
| CN113077420A (en) * | 2021-03-19 | 2021-07-06 | 江南大学 | Fish fiber evaluation method and system based on convolutional neural network |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR3071393B1 (en) * | 2017-09-27 | 2019-09-20 | Seb S.A. | SYSTEM AND METHOD FOR COOKING RICE |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101271522A (en) * | 2008-05-21 | 2008-09-24 | 中国农业大学 | Automatic recognition method for yellow-colored rice in rice |
| CN101819141A (en) * | 2010-04-28 | 2010-09-01 | 中国科学院半导体研究所 | Maize variety identification method based on near infrared spectrum and information processing |
-
2011
- 2011-12-31 CN CN201110457088.6A patent/CN102564993B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101271522A (en) * | 2008-05-21 | 2008-09-24 | 中国农业大学 | Automatic recognition method for yellow-colored rice in rice |
| CN101819141A (en) * | 2010-04-28 | 2010-09-01 | 中国科学院半导体研究所 | Maize variety identification method based on near infrared spectrum and information processing |
Non-Patent Citations (2)
| Title |
|---|
| 吴婧: "丹参白芍的红外光谱研究", 《中国优秀硕士学位论文全文数据库（医药卫生科技辑）》 * |
| 马冬红等: "近红外光谱技术在食品产地溯源中的研究进展", 《光谱学与光谱分析》 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103837492B (en) * | 2014-02-24 | 2015-11-25 | 西北农林科技大学 | A kind of Kiwi berry based on near-infrared spectrum technique expand fruit lossless detection method |
| CN103837492A (en) * | 2014-02-24 | 2014-06-04 | 西北农林科技大学 | Nondestructive detection method for swelled kiwi fruits based on near-infrared spectrum technology |
| CN104198428A (en) * | 2014-08-21 | 2014-12-10 | 中国农业大学 | Method and system for rapidly identifying authenticity of seeds with seed coatings |
| CN104198428B (en) * | 2014-08-21 | 2016-08-24 | 中国农业大学 | Band seed coat agent seed authenticity rapid identification method and system |
| CN104215584B (en) * | 2014-08-29 | 2017-01-04 | 华南理工大学 | A kind of detection method distinguishing the rice place of production based on hyper-spectral image technique |
| CN104215584A (en) * | 2014-08-29 | 2014-12-17 | 华南理工大学 | Hyper-spectral image technology-based detection method for distinguishing rice growing areas |
| CN105021562A (en) * | 2015-07-06 | 2015-11-04 | 黑龙江八一农垦大学 | Traceable method for rice origin and application thereof |
| CN105738340A (en) * | 2015-11-05 | 2016-07-06 | 新疆大学 | Quick detection method of lavender essential oil categories on the basis of Fourier-Raman spectrum |
| CN105678632A (en) * | 2016-03-15 | 2016-06-15 | 江西融思科技有限公司 | Method and structure for tracing detonator manufacturer according to detonator tube shell |
| CN105675534A (en) * | 2016-03-25 | 2016-06-15 | 北京市农林科学院 | Method for quickly and nondestructively identifying polished grains |
| CN106092822A (en) * | 2016-06-07 | 2016-11-09 | 河南工业大学 | A kind of distinguish waxy with non-waxy rice method |
| CN106198442A (en) * | 2016-09-14 | 2016-12-07 | 南京财经大学 | A kind of spectroscopic analysis methods differentiating Panjin rice |
| CN110118748A (en) * | 2019-04-29 | 2019-08-13 | 山西农业大学 | The construction method of the fourier-transform infrared finger-print of different ageing time Shanxi mature vinegars and application |
| CN111401794A (en) * | 2020-04-24 | 2020-07-10 | 江苏傲农生物科技有限公司 | Feed quality control method based on near infrared spectrum |
| CN112161949A (en) * | 2020-09-16 | 2021-01-01 | 贵州国台酒业股份有限公司 | Method for identifying Maotai-flavor liquor brewing process based on infrared spectrum technology |
| CN112161949B (en) * | 2020-09-16 | 2023-07-18 | 贵州国台酒业集团股份有限公司 | Method for identifying Maotai-flavor liquor brewing process based on infrared spectrum technology |
| CN112485216A (en) * | 2020-11-20 | 2021-03-12 | 华中农业大学 | Multi-source information fusion Thailand jasmine rice adulteration identification method |
| CN112485216B (en) * | 2020-11-20 | 2022-02-01 | 华中农业大学 | Multi-source information fusion Thailand jasmine rice adulteration identification method |
| CN113077420A (en) * | 2021-03-19 | 2021-07-06 | 江南大学 | Fish fiber evaluation method and system based on convolutional neural network |
| CN113077420B (en) * | 2021-03-19 | 2024-05-31 | 江南大学 | Fish fiber evaluation method and system based on convolutional neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102564993B (en) | 2015-07-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102564993B (en) | Method for identifying rice varieties by using Fourier transform infrared spectrum and application of method | |
| Wei et al. | Visual detection of the moisture content of tea leaves with hyperspectral imaging technology | |
| He et al. | Discrimination of varieties of tea using near infrared spectroscopy by principal component analysis and BP model | |
| CN111968080B (en) | Method for detecting quality of inside and outside of Feicheng peaches based on hyperspectral and deep learning | |
| CN110411957B (en) | Nondestructive rapid prediction method and device for shelf life and freshness of fruits | |
| CN104062257B (en) | A kind of based on the method for general flavone content near infrared ray solution | |
| CN103235095B (en) | Water-injected meat detection method and device | |
| CN102305772A (en) | Method for screening characteristic wavelength of near infrared spectrum features based on heredity kernel partial least square method | |
| CN108072626A (en) | A kind of pitch brand identification method | |
| Wang et al. | Rapid detection of protein content in rice based on Raman and near-infrared spectroscopy fusion strategy combined with characteristic wavelength selection | |
| CN104062258A (en) | Method for rapid determination of soluble solids in compound ass-hide glue pulp by near infrared spectroscopy | |
| CN105092519A (en) | Sample composition determination method based on increment partial least square method | |
| CN106198447A (en) | Chemical Mixed Fertilizer main component harmless quantitative detection method based on near-infrared spectrum technique | |
| CN108760677A (en) | A kind of rhizoma pinellinae praeparata based on near-infrared spectrum technique mixes pseudo- discrimination method | |
| Zhao et al. | Hyperspectral imaging coupled with random frog and calibration models for assessment of total soluble solids in mulberries | |
| Wu et al. | Variety identification of Chinese cabbage seeds using visible and near-infrared spectroscopy | |
| CN101957311A (en) | Qualitative analysis method in tablet drying process | |
| CN113655027A (en) | Method for rapidly detecting tannin content in plant by near infrared | |
| CN108645809A (en) | A kind of method that near-infrared spectrum technique quickly identifies rosin original tree species | |
| CN105784629B (en) | The method that the stable carbon isotope ratio of soil is quickly detected using middle infrared spectrum | |
| Sun et al. | Nondestructive identification of barley seeds varieties using hyperspectral data from two sides of barley seeds | |
| Song et al. | Rapid identification of adulterated rice based on data fusion of near-infrared spectroscopy and machine vision | |
| CN109932335A (en) | It is a kind of for the method for natural rubber assay in plant and measurement use LED near infrared spectrometer | |
| CN104181125A (en) | Method for rapidly determining Kol-bach value of beer malt | |
| Zhang et al. | Rapid non-destructive identification of selenium-enriched millet based on hyperspectral imaging technology. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150715 Termination date: 20151231 |
|
| EXPY | Termination of patent right or utility model |