JP5274288B2 - Flour discrimination method and apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a grain flour discriminating method and apparatus which enable the easy and certain discrimination of the classification such as a kind or the like of grain flour without requiring cost, labor and time, and pretreatment. <P>SOLUTION: The grain flour discriminating method includes the step of acquiring the particle size of the grain flour MF becoming a measuring target, the step of acquiring the excitation/fluorescence matrix (EEM) data of the grain flour by measuring the intensity of the fluorescence produced from the grain flour irradiated with exciting light while stepwise changing an exciting wavelength and a fluorescence wavelength within respective predetermined wavelength ranges, the step of parameterizing the fluorescent characteristics of the grain flour by the statistic analyzing processing of the EEM data to extract the same, the step of preliminarily acquiring the parameterized fluorescent characteristics of the particle size and classification of known grain flour to store the same in a data base, and the step of discriminating the classification of the grain flour becoming the measuring target by collating the extracted fluorescent characteristics of the grain flower with the parameterized fluorescent characteristics of the known grain flour, which are preliminarily obtained corresponding to the classification and particle size of the known grain flour, using the obtained particle size. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a method and apparatus for determining flour, a quality control method for flour products, and flour products. More specifically, the present invention relates to flour discrimination in which the type of flour is identified using excitation / fluorescence matrix (EEM) measurement. The present invention relates to a method, a flour discrimination device to which the method is applied, a quality control method of a flour product using the flour discrimination method, and a flour product quality-controlled by this.

  2. Description of the Related Art Conventionally, components in foods have a great influence on food taste, texture, safety, and the like, and therefore, specific components contained in foods have been extracted and identified by chemical analysis or the like. Especially for agricultural products and their processed products, such as flour and processed foods, the varieties are set according to the quality, and the use and price are determined according to the varieties. It is important to detect components and their contents for properly maintaining the components and their contents included depending on the variety. In addition, these agricultural products and their processed foods may use various chemical substances such as agricultural chemicals and preservatives in the production, distribution, and processing stages, and their effects on the human body have been pointed out. Therefore, the residue of chemical substances such as agricultural chemicals and the generation of bacteria are severely restricted by the Food Sanitation Law and the Agricultural Chemicals Control Law, etc., so detection of these chemical substances, detection of their contents, and radiation exposure Detection etc. are strictly performed.

However, detection of each component in food, pesticide residue, etc. and detection of its content are carried out by chemical analysis, etc., so it takes time and labor, and part or all of the food must be chemically treated. Therefore, only a part of the food or the whole food can be detected as an average value, and in the case of detection of a part of the food, it can only be detected as being uniform in the whole food. It cannot be inspected in a state that is non-destructive.
In contrast, in recent years in foods that have been sliced with a microtome, etc., by combining spectral spectrum measurement and excitation-fluorescence matrix (EEM) measurement with two-dimensional imaging and three-dimensional graphic imaging technologies. The component distribution, the internal structure of food, etc. are visualized so that specific component information, component distribution information, form information, safety information, etc. inside the food can be easily observed visually.

For this reason, the present inventors disclosed in Patent Document 1 a plurality of objects on the measurement object while changing the excitation wavelength to be irradiated and the fluorescence wavelength to be observed in a stepwise manner within a predetermined excitation wavelength range and a predetermined fluorescence wavelength range. Excitation / fluorescence matrix information acquisition process for acquiring excitation / fluorescence matrix (hereinafter also simply referred to as EEM) information of each measurement location on the measurement object by measuring the fluorescence intensity at each measurement location, for each measurement location The acquired EEM information is compressed by statistical analysis processing to extract the fluorescence characteristics of each measurement location, and the color determination for determining the color to be colored at each measurement location based on the compressed EEM information Component distribution visible including a component distribution visualization step that visualizes the component distribution of the measurement object by color mapping the color to be colored at each measurement location It has proposed a method.
In the above-mentioned Patent Document 1, the visualization method acquires component distribution information in the cross section of the measurement object, and also visualizes the component distribution, thereby analyzing the internal structure of the measurement object, detecting residual agricultural chemicals, and radiation irradiation. It can be used for such as.

Japanese Patent No. 3706914

By the way, when the component distribution visualization method disclosed in Patent Document 1 is used, it is true that the component distribution information of the internal structure can be easily manipulated in the sliced cross section of the measurement object without chemically treating the measurement object. And a plurality of specific components can be simultaneously displayed and visualized.
However, in the component distribution visualization method of Patent Document 1, since it is necessary to detect the distribution of the sliced cross section of the measurement object, a plurality of excitation wavelengths and a plurality of fluorescences are present at all measurement locations existing on the sliced cross section. Not only is it necessary to acquire a large amount of EEM information for a combination of wavelengths, but it is also necessary to perform statistical analysis processing on the acquired large amount of EEM information to determine the colors to be colored for all measurement locations. There is a problem that it takes a long time to display and visualize the component distribution. For this reason, the algorithm to be applied is limited to an algorithm whose final purpose is visualization and has as little computation as possible.

On the other hand, in the varieties of flour products made from cereal milling, in particular in wheat product varieties made from wheat milling, Japanese and Western confectionery can be used according to, for example, thin flour, medium flour, strong flour, durum flour (durum semolina), etc. Main applications such as yam tempura, noodles, breads, macaroni and spaghetti have been determined, and it is necessary to strictly adjust the ingredients.
For this reason, the quality of wheat, which is a raw material for milling, is often greatly different depending on the production period, production location, weather conditions, storage conditions after harvesting, varieties, and the like. As a result of analyzing the quality (specific components) of the final product of the flour obtained by performing a series of milling operations using a portion of the sample as a sample, as well as the raw wheat varieties, The quality is also judged, and the actual production of the flour as a product is performed based on the result. Moreover, in the production of flour such as wheat flour, intermediate products (powder) and final products produced in various stages and processes of milling are extracted during production according to the product varieties. The type, variety (mixed state of components) or components are detected by the above-described chemical analysis or the like, and the subsequent milling operation is performed based on the detection result.

However, these chemical methods often require pretreatment, and as described above, there are problems of cost, labor, and time, and a sample of each raw wheat (product) is prepared in advance. In addition, there is a problem that it takes more cost, labor and time when it is subjected to the milling operation, and as a result, the analysis result cannot be reflected in the milling operation until the result of the chemical analysis is obtained. .
For this reason, at the flour milling site such as wheat flour, detection of raw wheat, intermediate products (powder), final products, etc. using chemical analysis as described above, use of new raw wheat, production of new final products, etc. In general, an expert visually determines the state of the intermediate product (powder) generated in various stages and processes of milling, and is used for milling operations such as mixing of the intermediate product (powder). The current situation is that adjustments have been made.
For this reason, in the production of such flour such as wheat flour, it is generated in various stages and processes of milling in order to perform adjustment according to the product type during manufacture without relying on adjustment by a skilled worker. There is a need for a method for easily detecting intermediate products (powder), final products, etc. during milling in a short time, preferably in real time.

However, since the component distribution visualization method of Patent Document 1 enables visualization of a multi-component component distribution, it is possible to visualize the component distribution of the cross-section of wheat flour. ) And final product varieties and ingredients are not easy to identify, and even if these varieties and ingredients can be identified by visualizing the component distribution, it takes a long time to visualize, so manufacturing There is a problem that it cannot be applied to identification of the above-mentioned varieties and components, and cannot be applied at a production site such as a mill where a short-time measurement is required.
In addition, the component distribution visualization method of Patent Document 1 acquires EEM information at a large number of measurement points when the physical condition of a measurement object such as thickness is constant, such as thinned wheat having a constant thickness. Therefore, it is possible to visualize the component distribution according to the EEM information by coloring, but when the measurement object is flour such as wheat flour, the acquired EEM information varies depending on the particle size of the flour. There was a problem that it was not possible to determine the type of sufficient flour varieties only by the difference in information.

Furthermore, in the past, when allergens are contained in food (in products), it has been taken up very often as a social problem because it causes various health hazards (food allergies) such as atopy. At present, the detection of allergens in foods also depends on chemical analysis and the like, and there is a problem that it takes cost, labor and time.
In order to prevent such food allergies, it is required to acquire information for determining whether the food contains allergen without cost, labor, and time.

In view of the above-mentioned problems of the prior art, the present invention eliminates cost, labor, and time, and is based on the type of flour, for example, the final product varieties of flour and intermediate products produced during the production of flour ( The main object of the present invention is to provide a flour discriminating method and apparatus capable of easily and reliably discriminating the types of flour) and varieties containing specific components without requiring pretreatment.
In addition to the main purpose described above, another object of the present invention is to determine the type of flour being produced at the flour production site. In the flour production process, the quality of products such as final products and the It is an object of the present invention to provide a flour discrimination method and apparatus that can be reflected in real time on chemical component information such as a main component, that is, adjustment of a milling operation according to the product type.
Another object of the present invention is to provide a flour quality control method and a flour product having a predetermined quality that can be determined to be a grain product having a predetermined quality.
Furthermore, in addition to the main object, another object of the present invention is to provide a flour discrimination method and apparatus capable of easily and reliably specifying the type of flour containing allergen and the allergen contained in the flour. It is in.
Another object of the present invention is to provide a quality control method for flour capable of determining that it is a cereal product having a predetermined quality that does not contain an allergen, and a flour product having a predetermined quality that does not contain an allergen. There is.

  In order to solve the above-mentioned problems, the present inventors have conducted intensive research on identifying types of flour varieties easily, reliably, and in real time without cost, labor and time. As a result, while the excitation wavelength and the fluorescence wavelength are changed stepwise in the predetermined excitation wavelength range and the predetermined fluorescence wavelength range, the excitation light is irradiated to the flour to be measured, and the fluorescence of the fluorescence generated from the measurement target flour It is parameterized by measuring the intensity to obtain excitation / fluorescence matrix (EEM) information, and subjecting the obtained EEM information to various statistical analysis processes such as principal component analysis to calculate parameterized fluorescence characteristics. That the fluorescence characteristics can be used as a parameter representing a fluorescent fingerprint peculiar to the measurement target flour of a predetermined particle size. Knowing that the flour to be measured can be separated, identified or discriminated by determining the parameterized fluorescence characteristics, and further, the particle size of the flour of a large number or the required number of known varieties and parameterized accordingly. The fluorescence characteristics (reference fluorescence characteristics) are obtained, and the known types of flour and the obtained particle sizes are associated with the parameterized fluorescence properties in advance in a database, and the particle sizes and parameterization of the flour to be measured are parameterized. Using the fluorescence characteristics, search the database, and depending on the particle size of the flour to be measured, parameterized fluorescence characteristics of the flour to be measured, and the grain size of the flour of known varieties obtained in advance By comparing with the reference fluorescence characteristics, the type of flour to be measured can be identified and specified, and the parameters of particle size and fluorescence characteristics Runode, and it found that easily match, and have reached the present invention.

  That is, the present invention provides, as a first aspect, a flour discrimination method for discriminating the type of flour to be measured according to predetermined characteristics of flour obtained by pulverizing grain, wherein the flour The particle size acquisition step of acquiring the particle size of the above, the excitation light was irradiated while stepwise changing the excitation wavelength irradiated to the flour and the fluorescence wavelength to be measured in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range Excitation / fluorescence matrix information acquisition step of measuring fluorescence intensity of fluorescence generated from flour and acquiring excitation / fluorescence matrix information of the flour, and excitation / fluorescence matrix information of the flour acquired in this information acquisition step Fluorescence characteristic extraction step for performing statistical analysis and parameterizing and extracting the fluorescence characteristic of the flour, and for the known types of flour, the particle size acquisition step, excitation / fluorescence in advance A trick information acquisition step and a fluorescence property extraction step are performed to obtain the particle size and the parameterized fluorescence property, the known type of the flour and the obtained particle size, and the parameterized fluorescence. The reference fluorescence characteristic acquisition / storage process that associates characteristics with the database and stores them in the database in advance, and the particle size of the flour to be measured acquired in the particle size acquisition process and the measurement extracted in the fluorescence characteristic extraction process Using the parameterized fluorescence characteristics of the target flour, search in the database, and according to the particle size of the measurement target flour, the parameterized fluorescence characteristics of the measurement target flour And the parameterized fluorescence characteristics of the flour for the known type of flour previously stored in the database. , There is provided a method for discriminating flour, which comprises the fluorescence properties matching step of determining the type of the flour to be the measurement target.

Here, it is preferable that the fluorescence characteristic extraction step acquires the fluorescence characteristic of the flour as a two-dimensional parameter from the excitation / fluorescence matrix information.
Moreover, it is preferable that the said fluorescence characteristic extraction process performs the said statistical analysis process by multivariate analysis or data mining.
The multivariate analysis or the data mining is preferably performed using at least one method selected from the group consisting of data structure analysis, discriminant analysis, pattern classification, multi-factor data analysis, regression analysis, and a learning machine.
The data structure analysis is preferably performed using at least one method selected from the group consisting of principal component analysis, factor analysis, correspondence analysis, and independent component analysis.

The fluorescence characteristic extraction step performs the multivariate analysis or the data mining using the principal component analysis method, and uses the excitation / fluorescence matrix information as a parameter with two or less principal component scores. It is preferable to quickly acquire them as fluorescent properties of the flour.
The discriminant analysis is preferably performed using a linear discriminant analysis method or a nonlinear discriminant analysis method.
Further, the fluorescence characteristic extraction step performs the multivariate analysis or the data mining using the discriminant analysis technique, and applies the excitation / fluorescence matrix information to a discriminant stored in the database in advance. It is preferable that the parameters are calculated by the above and compared with the parameters stored in advance in the database to determine the type of flour.
The pattern classification is preferably performed by cluster analysis or multidimensional scaling.
The regression analysis is preferably performed using a linear regression, nonlinear regression, or multiple regression analysis technique.

In the fluorescence characteristic extraction step, the multivariate analysis or the data mining is performed using the method of the multiple regression analysis, and the excitation / fluorescence matrix information is applied to a calibration curve stored in the database in advance. It is preferable to calculate parameters and to collate them with parameters stored in advance in the database.
Further, the learning machine is preferably performed using at least one method selected from the group consisting of a neural network, a self-organizing map, group learning, and a genetic algorithm.
In addition, the flour is flour containing a specific component of the grain as a main component, and the fluorescence characteristic matching step includes the fluorescence property of the flour extracted in the fluorescence property extraction step and each component of the flour. In accordance with the flour and its particle size, it is obtained in advance and stored in the database, and collated with the parameterized fluorescence characteristics of the flour containing each component of the grain as a main component, and the flour to be measured Is preferably a flour containing a specific component of the grain as a main component.

Moreover, it is preferable that the said grain flour contains at least 1 selected from the group which consists of wheat flour, barley flour, rye flour, rice flour, corn flour, buckwheat flour, beans flour, and mixtures thereof.
Moreover, it is preferable that the said wheat flour contains at least 1 selected from the group which consists of durum wheat flour, strong flour, medium flour, and thin flour.
Moreover, it is preferable that the flour used as the measurement object is flour containing an allergen therein, and the type of the flour used as the measurement object and the allergen contained in the flour are specified.
Moreover, it is preferable that the said flour used as the said measuring object is a mixture of many types of flour, and the kind of the said flour used as the measuring object and the specific flour contained in this flour are specified.
Moreover, it is preferable that the said particle size acquisition process acquires the known particle size measured beforehand of the said flour used as the said measuring object as information.

  Moreover, in order to solve the said subject, this invention is a 2nd aspect. WHEREIN: According to the predetermined characteristic which the flour obtained by grind | pulverizing the grain determines the type of the flour to be measured. A discrimination device, a particle size acquisition unit that acquires the particle size of the flour, a spectral illumination unit that irradiates the flour with the excitation light while changing the excitation wavelength stepwise in a predetermined excitation wavelength range, and The excitation light is irradiated while steppingly changing the fluorescence wavelength to be measured in a predetermined fluorescence wavelength range according to the stepwise change in the excitation wavelength of the excitation light irradiated onto the flour by the spectral illumination means. Information acquisition means for measuring fluorescence intensity of fluorescence generated from the applied flour and obtaining excitation / fluorescence matrix information of the flour, and excitation / fluorescence matrix information of the flour obtained by the information acquisition means A fluorescence characteristic extraction means for performing a statistical analysis process to parameterize and extract the fluorescence characteristics of the flour, and a flour whose type and particle size are known in advance, the spectral illumination means, the information acquisition means, and the fluorescence characteristic extraction A database for storing in advance the parameterized fluorescence characteristics of flour with known type and particle size obtained using means and the known type and particle size of the flour in association with each other, and the measurement Using the known particle size of the target flour and the parameterized fluorescence characteristics of the measurement target flour extracted by the fluorescence characteristic extraction means, the database is searched, and the flour of the measurement target The parameterized fluorescence characteristics and the granularity corresponding to the known granularity of the flour to be measured, stored in advance in the database Flour further comprising: a fluorescence characteristic collating means for collating with the parameterized fluorescence characteristic of the flour for the known type of the flour to determine the type of the flour to be measured A discriminating apparatus is provided.

  Moreover, in order to solve the said subject, this invention discriminate | determines that the 3rd aspect is a flour product with predetermined | prescribed quality using the said discrimination method or discrimination device of flour. It provides a quality control method for products.

According to the flour discriminating method and apparatus of the present invention, the type of flour, for example, the varieties of the final product of flour, or the production of flour, without cost, labor and time, and without requiring pretreatment It is possible to easily and reliably discriminate the types of intermediate products (powder) produced in the past and the types such as varieties containing specific components.
Further, according to the method and apparatus for determining flour of the present invention, in addition to the above effects, the type of flour including physical information such as chemical component information and particle size of the flour being produced at the flour production site. In the flour production process, it is possible to reflect in real time on the quality of the final product, etc., and the chemical component information such as its main component, that is, the adjustment of the milling operation according to the product type it can.
Therefore, according to the quality control method and flour product of the flour of the present invention, it is possible to manage that the product is a grain product having a predetermined quality by using the above-described flour discrimination method or discrimination device. A flour product with quality can be provided.

Furthermore, according to the flour discriminating method and apparatus of the present invention, in addition to the above effects, the type of flour containing allergen and the allergen contained in the flour can be easily and reliably specified.
As a result, according to the quality control method and product of flour of the present invention, it can be managed that it is a grain product having a predetermined quality that does not contain allergens. can do.

  BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, a flour discrimination method according to the present invention, a flour discrimination device that implements the discrimination method, a quality control method of flour using the discrimination method or the discrimination device, and a product quality-controlled by this are shown in the accompanying drawings. This will be described in detail based on such embodiments.

[1] Method for discriminating flour First, a method for discriminating the type of flour of the present invention will be described with reference to FIGS. 1 and 2. The flour discrimination method of the present invention is not limited to this.

In FIG. 1, an example of the flowchart of the discrimination method of the flour of this invention is shown.
The flour discrimination method of the present invention is a flour discrimination method for discriminating the type of flour according to the predetermined characteristics of flour obtained by pulverizing the grain. Acquires the parameterized fluorescence characteristics of flour according to the type of each flour and its particle size as the reference fluorescence characteristics for the particle size, and associates the known types and particle sizes with the parameterized fluorescence properties and stores them in the database in advance The reference fluorescence characteristic acquisition / storage step (step S10) to be performed, the particle size acquisition step (step S12) to acquire the particle size of the flour to be measured, and the measurement target in the predetermined excitation wavelength range and the predetermined fluorescence wavelength range Excitation and fluorescence of flour by measuring the fluorescence intensity of the fluorescence generated from flour irradiated with excitation light while gradually changing the excitation wavelength and the fluorescence wavelength to be measured. EEM information acquisition process (step S14) for acquiring tricks (Excitation-Emission Matrix) information, and EEM information of flour obtained in the EEM information acquisition process of step S14 is subjected to statistical analysis processing, and fluorescence characteristics of flour are parameterized Fluorescence characteristics of flour extracted in the fluorescence characteristic extraction process of step S16 using the particle size obtained in the fluorescence characteristic extraction process (step S16) and the particle size acquisition process of step S12, and reference to step S10 Including a fluorescence characteristic collation step (step S18) for collating the reference fluorescence characteristic corresponding to the grain size of the flour obtained in advance in the fluorescence characteristic acquisition / storage process to determine the type of flour to be measured. It is a feature.

Here, the excitation / fluorescence matrix (EEM) is a contour graph (fluorescence fingerprint) composed of three axes of excitation wavelength λ _Ex , fluorescence wavelength (measurement wavelength) λ _Em , fluorescence intensity I _{Ex, and Em} , and flour The peculiar pattern of the containing component of is shown.
As an example of the EEM, for example, an EEM of pasta powder (durum powder) of Example 1 described later is shown in FIG.
As shown in FIG. 6 (a), the EEM can be represented in a plane by plotting the intensity of fluorescence at each point by color and plotting contour lines with the horizontal axis as the excitation wavelength and the vertical axis as the fluorescence wavelength. EEM has the advantages of being able to characterize the sample without pre-processing, being easy to operate and measuring in a short time, and having higher sensitivity compared to ultraviolet absorbance, so it can identify seawater contaminants. It is also a technique that is widely used for specifying raw materials for dyeing. The present invention is a new technique for identifying flour from the EEM as a target by finding target information from the EEM through statistical analysis processing.
Here, in the case of flour in which various components that are the subject of the present invention are mixed, it is considered that the pattern in which the EEM of each component is superimposed is formed. For this reason, in the present invention, a unique EEM can be obtained according to the type of flour varieties with different mixing states of various components. Therefore, the type of flour is to be determined using this unique EEM. It is.

In the present invention, if any flour is obtained by pulverizing cereals, any components and substances (allergens, chemical substances, agricultural chemicals, impurities, preservatives, etc.) are mixed in the flour. In principle, all flours can be measured.
Any molecule has an absorption / fluorescence pattern corresponding to its molecular structure, and even if it does not emit any fluorescence, there is information that “EEM that does not emit fluorescence”, and it can be distinguished from other substances based on that information. It is.
In addition, the flour used as a measuring object may contain the water | moisture content. Since the fluorescence intensity of water is low, it can be measured without being affected by moisture compared to the case of measurement using infrared (IR) or near infrared (NIR).

Here, examples of the flour to be measured include wheat flour, barley flour, rye flour, rice flour, corn flour, oat flour, beans flour, and mixtures thereof.
Examples of wheat flour include durum flour (durum semolina), strong flour, medium flour, and weak flour.
Examples of allergens and contaminants contained in the flour to be measured include soybeans, rapeseed, milo, maize, buckwheat kazura, eggs, milk / dairy products, buckwheat and peanuts.
Moreover, agrochemicals, preservatives, etc. are mentioned as a chemical substance contained in the flour used as a measuring object.

First, the particle size acquisition process which acquires the particle size of the flour used as the measuring object of step S12 is demonstrated.
In the present invention, the particle size acquisition step may be acquired by any method as long as the particle size of the flour to be measured can be acquired, and the particle size of flour already measured by a particle size measurement method or apparatus is known. May be acquired as particle size information, or may be acquired by measuring the particle size of the flour to be measured by a particle size measurement method or apparatus in the particle size acquisition step. In addition, when acquiring the known particle size of flour as information, even if it is the particle size measured in advance in the implementation of the flour discrimination method of the present invention, it is previously determined regardless of the implementation of the flour discrimination method of the present invention. It may be a particle size measured and given to the flour as particle size information.
In the present invention, the reason for obtaining the grain size of the flour is that the EEM information obtained by the difference in the grain size is different even for the same variety of flour (the same chemical components). That is, in order to accurately determine the type of flour variety, it is necessary to consider the difference in EEM information due to the difference in particle size.

Here, the particle size measurement method of the flour is not particularly limited, but any method may be used as long as the particle size of the flour can be measured, including various conventionally known flour particle size measurement methods. A particle size acquisition method can be used. For example, a measuring method using a particle size distribution measuring apparatus by a laser diffraction / scattering method such as Microtrac (manufactured by Nikkiso Co., Ltd.), Mastersizer (manufactured by Malvern Co., Ltd.) or the like is preferably used.
In addition, as long as the particle size acquisition process of step S12 is before the fluorescence characteristic collation process of step S18, you may perform it at any time. Further, when the particle size of the flour to be measured is known, as described above, the particle size of the flour to be measured may be acquired as information using the known particle size without performing the particle size measurement or the like.

In addition, what is necessary is just to obtain and convert into the average particle diameter as a particle size of flour, and it is not restrictive in particular, What particle size may be sufficient.
In addition, it is preferable that the particle size distribution of the flour is as small and sharp as possible, and the dispersion is preferably small with respect to the average particle size of the flour. The reason is that since the type of flour is discriminated, it is preferable that there is less variation in the particle size of the flour that affects the EEM information used for discriminating the type. This is because it is possible to distinguish flour even if only one point of flour is measured without consideration. If it is one-point measurement, it can also be used at the manufacturing site where measurement is required in a short time.
In addition, when the particle size distribution of the flour is widened, so-called broad, the EEM information of the flour is measured at a plurality of points, and they may be used on average. It is good also as a measuring object in this invention after aligning.

Next, the EEM information acquisition process which acquires the EEM information of the flour used as a measuring object is demonstrated (step S14).
The EEM information of the flour to be measured is irradiated with excitation light while gradually changing the excitation wavelength λ _Ex and the fluorescence wavelength λ _Em to be measured in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range. It can be obtained by measuring fluorescence intensity (spectral information) I _{Ex, Em} of fluorescence generated from the processed flour.

Or EEM information becomes what dimensional data, what kind of excitation wavelength lambda _Ex and fluorescence wavelength lambda _Em combination with fluorescence intensity I _{Ex of,} depends on how to measure the _Em. For example, m-dimensional EEM information can be acquired by measuring the fluorescence intensity I _{Ex, Em} with a combination of one type of excitation wavelength λ _Ex and m types of fluorescence wavelengths λ _Em (1 × m = m).
The number m of the combination of the excitation wavelength λ _Ex and the fluorescence wavelength λ _Em is a positive integer and is preferably an integer of 5 or more. However, when m is small, the amount of EEM information obtained is small and an accurate component Therefore, m is more preferably 50 or more, and even more preferably 100 or more. Further, if m is too large, the subsequent statistical processing becomes complicated, so m is preferably 1000 or less, more preferably 500 or less.

In FIG. 2, in order to obtain EEM information of flour while changing the excitation wavelength and the fluorescence wavelength stepwise in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range, the fluorescence intensity of the flour to be measured is measured. An example of the flowchart of the fluorescence intensity measurement process to measure is shown.
Excitation wavelength range, excitation / fluorescence slit width (excitation wavelength pitch), fluorescence wavelength range, excitation / fluorescence acquisition interval (fluorescence wavelength pitch), and scan speed should be set appropriately by the measurer according to the flour to be measured. Can do.

As shown in FIG. 2, the flour as the measurement target is stored in the cell (step S20). While changing the excitation wavelength and the fluorescence wavelength step by step in the predetermined excitation wavelength range and the predetermined fluorescence wavelength range (steps S22 and S26), the fluorescence intensity (spectrum information) of the flour to be measured is measured (step) S30).
For example, when acquiring EEM information under the measurement conditions of [excitation wavelength range: 200 to 900 nm / 10 nm interval, fluorescence wavelength range: 200 to 900 nm / 10 nm interval], the excitation wavelength applied to the flour to be measured is set to 200 nm. Setting is performed (step S22), the fluorescence wavelength to be measured is increased by 10 nm by 200 nm, 210 nm,... (Step S26, S28), and the fluorescence intensity up to the measurement fluorescence wavelength of 900 nm is measured (step S30). When the measured fluorescence wavelength exceeds 900 nm (step S28), the excitation wavelength to be irradiated is then increased by one pitch from 200 nm to 210 nm (steps S22, S24), and the measured fluorescence wavelength is increased to 200 nm, 210 nm,. (Steps S26 and S28), and the fluorescence intensity up to the measurement fluorescence wavelength of 900 nm is measured (step S30). The measurement of the fluorescence intensity from the measurement fluorescence wavelength 200 nm to 900 nm (steps S26, S28, S30) is repeated by increasing the irradiation excitation wavelength from 200 nm to 900 nm by 10 nm (steps S22, S24). When the irradiation excitation wavelength exceeds 900 nm (step S24), the measurement of the flour fluorescence intensity is terminated.
Under this measurement condition, a total of 5041 fluorescence intensity measurement values are acquired. EEM information can be obtained by displaying the measurement values of the fluorescence intensity under the 5041 conditions obtained in this manner in a matrix display with respect to the excitation wavelength and the fluorescence wavelength.

EEM information of flour can be obtained by measuring the fluorescence intensity of flour as a measurement target while changing the excitation wavelength and fluorescence wavelength stepwise in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range. . Or EEM information becomes what dimensional data, what kind of excitation wavelength lambda _Ex and fluorescence wavelength lambda _Em combination with fluorescence intensity I _{Ex of,} depends on how to measure the _Em. That is, when the number of fluorescence intensity measurements is m, m-dimensional EEM information can be acquired.
In the example of the above measurement conditions, the excitation wavelength range is 200 to 900 nm, the fluorescence wavelength range is 200 to 900 nm, and the total of I ^M _{Ex200, Em200} , I ^M _{Ex200, Em210} , ... I ^M _{Ex890, Em900} , I ^M _{Ex900, Em900} Since the fluorescence intensity measurement value consisting of the parameters of the 5041 condition is acquired, it has 5041-dimensional EEM information.

In addition, since the fluorescence wavelength which the flour which is a measuring object emits is longer than an excitation wavelength, you may measure a fluorescence intensity with the fluorescence wavelength longer than an excitation wavelength.
For example, when the irradiation excitation wavelength is set to 500 nm, since no fluorescence is observed when the measurement fluorescence wavelength is less than 500 nm, it is not necessary to measure the fluorescence intensity in the measurement fluorescence wavelength range of 200 to 490 nm, but in the range of about 500 to 900 nm. It is sufficient to measure the fluorescence intensity.
Alternatively, from the measurement result of the fluorescence intensity of the fluorescence generated from the flour obtained under the above measurement conditions [excitation wavelength range 200 to 900 nm, fluorescence wavelength range 200 to 900 nm], the fluorescence intensity in the fluorescence wavelength range shorter than the excitation wavelength is calculated. It may be deleted.
In the fluorescence characteristic extraction process described later, when it becomes clear that the excitation wavelength and the fluorescence wavelength can be sufficiently discriminated within a certain limited range or a specific wavelength, only the measurement in the wavelength range or the wavelength is performed. Just do it.

The EEM information acquired in this way may be subjected to various known signal processing for the purpose of reducing noise information other than grain quality and removing disturbances.
Examples of signal processing methods include wavelet transform, Fourier transform, peak detection, smoothing, differentiation, spectrum correction, approximation / interpolation and waveform removal, Savitzky-Golay, moving average, normalization, and the like.

In the EEM information acquisition step of step S14, parameters I ^M _{Ex200, Em200} , I ^M _{Ex200, Em210} ,... I under 5041 conditions acquired under the above measurement conditions [excitation wavelength range 200 to 900 nm, fluorescence wavelength range 200 to 900 nm]. based on the _{^{M Ex890, Em900, I M Ex900}} , Em900, can be distributed to 5041 dimensional space EEM information. The distribution information in the 5041-dimensional space represents the type of flour flour (chemical component information) and the difference in particle size. However, since the distribution information in the 5041-dimensional space is high-dimensional, it cannot be used as it is for determining the type of flour.
Therefore, the EEM information of the flour to be measured is reduced to a lower dimension by statistical analysis processing, and parameterized so as to express the characteristics of the type of flour flour, preferably two-dimensional parameterized, and measured. Extract fluorescence characteristics of target flour.
Below, this fluorescence characteristic extraction process is demonstrated (step S16).

The statistical analysis process may be any parameterization, preferably two-dimensional parameterization so as to express the characteristics of the type of flour variety, but is preferably performed by multivariate analysis or data mining.
Specific examples thereof include data structure analysis, discriminant analysis, pattern classification, multi-factor data analysis, regression analysis, and learning machine.

Data structure analysis includes principal component analysis, factor analysis, correspondence analysis, and independent component analysis.
Further, as the discriminant analysis, linear discriminant analysis or non-linear discriminant analysis can be cited.
Here, the canonical discriminant analysis is an example of the linear discriminant analysis, and the decision tree is an example of the nonlinear discriminant analysis.
Examples of pattern classification include cluster analysis and multidimensional scaling.
Examples of regression analysis include linear regression and nonlinear regression.
Here, examples of the linear discriminant analysis include Partial Last Square (PLS) regression, single regression analysis, multiple regression analysis, and principal component regression, and examples of the nonlinear discriminant analysis include logistic regression and regression trees.
Examples of learning machines include neural networks, self-organizing maps, group learning, and genetic algorithms.
Any analysis method may be used for the statistical analysis process as long as each difference (type difference) of the varieties of flour to be measured can be analyzed most accurately.

In the fluorescence characteristic extraction step of step S16, it is preferable that m-dimensional EEM information is acquired as a two-dimensional parameter. By obtaining EEM information as a two-dimensional parameter from m-dimension, while maintaining the characteristics of EEM information, the grain size of the flour to be measured and the difference in EEM in the variety (chemical component) are measured. It can be acquired as a simple parameter with which the type of product can be identified.
For example, principal component analysis can be performed on the multidimensional EEM information obtained in the EEM information acquisition step of step S14, and two principal component scores 1 and 2 can be obtained as two-dimensional parameters. The two principal component scores represent the types of flour varieties to be measured at a predetermined particle size. That is, in the fluorescence characteristic extraction process of step S16, when performing principal component analysis as statistical analysis processing, it is preferable to obtain a two-dimensional parameter composed of two principal component scores 1 and 2 as parameterized fluorescence characteristics.

Next, a description will be given of a reference fluorescence characteristic acquisition / storage process in which parameterized fluorescence characteristics are calculated in advance according to the grain size and type of varieties of known flours to be verified and stored in a database (step) S10).
In the present invention, the EEM information acquisition step in step S14 and the fluorescence characteristic extraction step in step S16 described above are performed in exactly the same manner on the flour to be collated whose grain size and variety type are known. In addition, it is necessary to obtain parameterized fluorescence characteristics as reference fluorescence characteristics.
In the case of measuring all flours, it is necessary to obtain the reference fluorescence characteristics for all flour varieties and particle sizes to be collated, so for many flour types (types of flour), the particle size It is necessary to obtain a reference fluorescence characteristic for each and store (store) it in a data storage device such as a database for collation. In addition, since it is parameterized as a reference fluorescence characteristic and the amount of data is limited, the reference fluorescence characteristic and particle size data for flour types that need to be obtained and stored in advance are It can be easily stored in a data storage device such as a database.

If the type of flour to be verified is unknown, or after extracting reference fluorescence characteristics, obtain chemical component information by chemical analysis, etc., and specify the type of flour. It is preferable to keep it. Moreover, when the particle size of flour is unknown, it is preferable to specify the particle size by performing the particle size acquisition process of step S12 mentioned above previously or after reference fluorescence characteristic extraction.
It should be noted that any of the three types of flour types, particle sizes, and parameterized fluorescence characteristics (reference fluorescence characteristics) to be verified may be obtained first and later, but at least the fluorescence characteristics of step S18. Prior to the matching step, at least the reference fluorescence characteristic, preferably all three characteristics, should be obtained.

Further, according to the knowledge of the present inventors, the reference fluorescence characteristics extracted in the fluorescence characteristic extraction step of step S16, for example, the principal component scores 1 and 2 extracted as the reference fluorescence characteristics by the principal component analysis are the reference fluorescence. Even if there is some variation in chemical component information and grain size within the type of flour varieties in spatial coordinates and plane coordinates with the number of parameters of the characteristic (parameterized fluorescence characteristics) as a dimension, a predetermined spatial region or plane It is known that it falls within the region (see, eg, FIGS. 8, 11 and 14).
Therefore, the reference fluorescence characteristic itself according to the type and grain size of the flour variety is not stored in the data storage device, but can be used when extracting the fluorescence characteristic of the unknown sample in the fluorescence characteristic extraction step of step S16. , General formula Y = f (x) (x: EEM information, Y: output information) created from EEM information obtained from known grain size and type, and reference according to type and grain size of flour Saving (storing) the spatial area in the spatial coordinates of the fluorescence characteristic and the planar area in the plane coordinates (distribution status of EEM information obtained from known grain size and type, etc.) in a data storage device. Anyway. By doing so, the amount of data stored in the data storage device can be reduced, and the verification in the fluorescence characteristic verification process of step S18 can be facilitated.

By the way, as a reference fluorescence characteristic to be compared with the fluorescence characteristic of the flour extracted in the fluorescence characteristic extraction step, parameterization of a known flour containing each component of the grain as a main component according to the known flour and its particle size. The obtained fluorescence characteristics are obtained and collated using these reference fluorescence characteristics, whereby the flour to be measured can be identified as flour containing a specific component of the grain as a main component. In addition, by obtaining the reference fluorescence characteristics for the types of known flour product varieties, the types of flour to be measured can be set as the types of flour product varieties.
When detecting whether the flour to be measured is flour containing allergens, the type of flour to be measured is specified by obtaining the reference fluorescence characteristics for known flours containing known specific allergens It can be determined that the flour contains allergens, and the allergen contained in the flour can be identified.

When mixed flour consisting of a mixture of multiple types of flour is the flour to be measured, the type of mixed flour to be measured is determined by obtaining the reference fluorescence characteristics for the known mixture of known multiple types of flour. The specific flour contained in this mixed flour can be identified.
In addition, when flour containing chemical substances such as agricultural chemicals is the flour to be measured, residual agricultural chemicals to be measured by obtaining reference fluorescence characteristics for known flour containing chemical substances such as known agricultural chemicals The type of flour containing chemical substances such as can be identified, and as a result, chemical substances such as residual agricultural chemicals contained in the flour can be identified.

Next, the fluorescence characteristic matching step will be described (step S18).
In the fluorescence characteristic collation process of step S18, using the particle size of the flour that is the measurement target obtained in the particle size acquisition process of step S12 and the fluorescence characteristic of the flour extracted in the fluorescence characteristic extraction process of step S16 as search keys, The database is searched, and the granularity and fluorescence characteristics that are the search keys are read from the database, and the granularity and reference fluorescence characteristics that match or substantially match or correspond to each other and the types of flours having these granularity and reference fluorescence characteristics are read and extracted. By identifying the fluorescence characteristics of the selected flour and the reference fluorescence characteristics of the read flour, the type of the flour to be measured is identified as the type of the flour to be verified that matches the fluorescence characteristics, and discriminated. can do. In addition, in the reference fluorescence characteristic acquisition and storage process of step S10, the reference fluorescence characteristics of flour obtained in advance are stored in the database according to the type of flour to be verified and the particle size thereof. As described above.

  That is, the fluorescence characteristics of the flour extracted in the fluorescence characteristic extraction step are parameterized for flour containing each component of the grain as a main component, which is obtained in advance according to the flour for each component of the grain and its particle size. By comparing with the fluorescence characteristics, it is possible to specify that the flour to be measured is flour containing a specific component of the grain as a main component. Since the comparison between the measured fluorescence characteristic and the reference fluorescence characteristic in the present invention is a comparison between parameters, it can be performed very easily and accurately.

In addition, when the particle sizes are the same or substantially the same, the method for collating the measured fluorescence characteristics with the reference fluorescence characteristics is not particularly limited. If it is more than a predetermined threshold value, the type of flour having the measured fluorescence characteristic can be specified as the type of flour having the reference fluorescence characteristic.
In addition, as described above, when the reference fluorescence characteristic to be verified is obtained as a spatial area of the parameter space coordinates or a planar area of the parameter plane coordinates, the measured fluorescence characteristics are represented by these spatial areas or Both can be collated depending on whether or not they are inside the plane area. Of course, when entering the space area or the plane area, the type of flour having the measured fluorescence characteristics can be specified as the type of flour having the space area or the plane area. .

According to the flour discrimination method of the present invention, when the flour to be measured is flour containing allergen, the type of flour to be measured can be judged to be flour containing a specific allergen, and this flour Allergens contained therein can be identified.
Further, according to the method for discriminating flour of the present invention, when the flour to be measured is a mixture of multiple types of flour, the type of the mixed flour to be measured can be specified and included in the mixed flour Specific flour can be identified.
Similarly, according to the flour discrimination method of the present invention, it is possible to specify chemical substances such as residual agricultural chemicals contained in the flour to be measured.
Note that the type of flour to be measured may be the type of flour product variety.
By the way, in the method for discriminating flour of the present invention, the type, particle size, and reference fluorescence characteristics of known flour varieties have been acquired in advance, so discrimination of flour products such as flour products and management of flour product manufacturing processes With the discrimination of intermediate production flour for, discrimination of flour consisting of each component of the grain or flour containing each component as a main component, discrimination of flour containing known allergens and other known components, and the discrimination result, It is suitable for identifying allergens and other components contained in flour.
The flour discrimination method of the present invention is basically configured as described above.

[II] Flour Discrimination Device Next, the flour discrimination device of the present invention that implements the flour discrimination method of the present invention will be described with reference to FIGS. The flour discrimination device of the present invention is not limited to the flour discrimination device in the illustrated example.
3 is a block diagram of an embodiment of the flour discrimination device of the present invention, and FIG. 4 is a block diagram of an embodiment of a measurement unit of the flour discrimination device shown in FIG.

As shown in FIG. 3, the flour discriminating apparatus 10 of the present invention is configured to measure the measurement target flour MF according to the predetermined characteristics of the measurement target flour (flour to be measured) MF obtained by pulverizing the grain. The type of quality is determined, and the measurement unit 12 that measures fluorescence intensity (spectrum information), the particle size acquisition unit 20 that acquires the particle size of the measurement target flour MF, and the type of quality of the measurement target flour MF are determined. And the particle size acquired by the measurement unit 12, the measurement result by the measurement unit 12, the measurement result by the measurement unit 12, the measurement operation by the measurement unit 12, the determination operation by the measurement unit 12 A display 18 for displaying a discrimination result by the flour discrimination unit 14, an operation instruction or input by the operation unit 16, and the like.
In addition, when the particle size of the measurement target flour MF is previously measured and obtained as information, the particle size of the measurement target flour MF is acquired by inputting the particle size of the measurement target flour MF as information through the operation unit 16. You may do it. In this case, since the input by the operation unit 16 is acquisition of the granularity, the granularity acquisition unit 20 may not be provided.

In addition, the flour discrimination device 10 of the present invention is a reference fluorescence characteristic (parameterized fluorescence) with known flour quality type and particle size, which is a target for collation for discriminating the quality type of the measurement target flour MF. The characteristic can also be used in advance. In this case, the flour whose type and particle size are known may be the measurement target flour MF.
In addition, the flour discrimination device 10 of the present invention is a target for collation for discriminating the quality type of the measurement target flour MF, and the particle size and reference fluorescence characteristics (parameterized fluorescence) of the flour of known quality type. Needless to say, it is also possible to obtain the characteristic) in advance. In this case, the flour whose type is known may be the measurement target flour MF.
In addition, in order to obtain the type of flour having the reference fluorescence characteristic in advance, it is sufficient to perform chemical analysis of the flour in advance, or conversely, a standard whose type and particle size are known in advance by chemical analysis or the like. You may use flour which becomes.

  Here, the measurement unit 12 measures and measures the fluorescence from the measurement target flour MF, and measures the measurement target flour support portion (hereinafter simply referred to as “measurement target flour support portion”) that positions and supports the measurement target flour MF at a predetermined measurement position. 21), a spectral illumination unit 22 that emits fluorescence from the measurement target flour MF by irradiating the measurement target flour MF positioned on the support unit 21 with excitation light of a predetermined wavelength, and the spectral illumination unit 22 A fluorescence measurement unit 24 that measures fluorescence of a predetermined fluorescence wavelength generated from the measurement target flour MF irradiated with excitation light of a predetermined wavelength and transmits a measured value of the measured fluorescence to the flour discrimination unit 14. Yes. The spectroscopic illumination unit 22 and the fluorescence measurement unit 24 constitute a spectroscopic measurement unit that performs the fluorescence intensity measurement step of the EEM information acquisition step (see FIG. 1) in step S14 of the above-described flour discrimination method of the present invention.

Further, the flour discrimination unit 14 acquires the excitation / fluorescence matrix (EEM) information of the measurement target flour MF from the fluorescence measurement value obtained by the fluorescence measurement unit 24, and the information acquisition unit 26 acquires the information. The EEM information of the measured flour MF to be measured is subjected to statistical analysis processing, and the fluorescence characteristic extraction unit 28 that extracts the parameterized fluorescence characteristics of the measurement target flour MF and the particle size obtained by the particle size acquisition unit 20 are used to obtain fluorescence. Fluorescence characteristics for discriminating the type of the measurement target flour MF by comparing the parameterized fluorescence characteristics of the measurement target flour MF extracted by the characteristic extraction unit 28 with the parameterized reference fluorescence characteristics obtained in advance The particle size of the measurement target flour MF acquired by the verification unit 30 and the particle size acquisition unit 20, the EEM information of the measurement target flour MF acquired by the information acquisition unit 26, and A memory (DB) 32 for storing the parameterized fluorescence characteristics of the measurement target flour MF extracted by the light characteristic extraction unit 28, and the types, particle sizes, reference fluorescence characteristics, etc. Particle size acquisition unit 20, spectral illumination unit 22 of measurement unit 12, and units of fluorescence measurement unit 24, information acquisition unit 26 of flour discrimination unit 14, fluorescence characteristic extraction unit 28, fluorescence characteristic verification unit 30 and memory (DB (database )) A control unit 34 that controls each unit 32 and also controls the entire discrimination device 10 of the present invention.
In addition to the memory (DB) 32, the flour discriminating unit 14 stores an external memory 36 for storing a number of flour types, particle sizes, reference fluorescence characteristics, and the like as comparison targets, as shown in FIG. Is preferably provided.

First, the particle size acquisition unit 20 will be described.
The particle size acquisition unit 20 is for acquiring the particle size of the measurement target flour MF. The particle size acquisition unit 20 may be anything as long as the particle size of the measurement target flour MF can be acquired in the particle size acquisition step (see FIG. 1) of step S12 of the flour discrimination method of the present invention described above. For example, a particle size information input device or a reading device for acquiring the particle size of the measurement target flour MF, which has been previously measured by a particle size measurement method or apparatus, as particle size information, or a measurement target A particle size measuring device for measuring the particle size of flour MF may be used. Note that the known particle size of the measurement target flour MF acquired as information by the particle size acquisition unit 20 may be a particle size measured in advance by the particle size measurement device of the particle size acquisition unit 20, or the determination of flour of the present invention. The particle size may be a particle size that is measured in advance by a particle size measurement device that is not related to the device 10 and that is given to the measurement target flour MF as particle size information. This granularity information may be described in a document attached to the measurement target flour MF, or may be stored as data in a recording medium, or the type of the measurement target flour MF in the memory (DB). It may be stored in association with.

Here, the particle size measurement device for flour used in the particle size acquisition unit 20 is not particularly limited, and any device may be used as long as it can measure the particle size of flour, including conventionally known particle size measurement devices for flour, Various known flour particle size acquisition devices can be used. For example, a particle size distribution measuring apparatus using a laser diffraction / scattering method such as Microtrac (manufactured by Nikkiso Co., Ltd.), Mastersizer (manufactured by Malvern Co., Ltd.) or the like is preferably used.
In addition, when the particle size of the measurement target flour MF is previously measured and known, as described above, the particle size acquisition unit 20 may not include the particle size measurement device, and the known particle size is input by the operation unit 16. In this case, it is needless to say that the particle size acquisition unit 20 itself is not necessary.

Next, each component of the measurement unit 12 will be described.
The measurement target flour support portion 21 (hereinafter, also simply referred to as the support portion 21) of the measurement unit 12 supports the measurement target flour MF by positioning it at a predetermined measurement position for spectroscopic measurement. The support stand which supports the cell which stores MF in a predetermined measurement position may be sufficient, or the support part 21 itself may be a powder fixing device which fixes the measurement target flour MF. In addition, as such a cell, the cell proposed in Japanese Patent Application No. 2007-215672 according to the applicant's application can be used, and the powder fixing device is one of the applicant's one. A fixing device disclosed in Japanese Patent Application Laid-Open No. 2002-228563, which is related to a human application, can be used.

Next, in order to generate fluorescence from the measurement target flour MF, the spectral illumination unit 22 changes the excitation wavelength stepwise in a predetermined excitation wavelength range, and emits excitation light having a predetermined wavelength to the measurement target flour MF. Irradiation. That is, the spectroscopic illumination part 22 can change arbitrarily the excitation wavelength of the excitation light irradiated to the measurement object flour MF.
As shown in FIG. 4, the spectral illumination unit 22 applies a light source 38 for emitting excitation light in a predetermined excitation wavelength range and excitation light having a predetermined excitation wavelength for irradiating the measurement target flour MF. A spectroscopic device 40 for spectroscopic analysis, an excitation wavelength adjusting unit 42 of the spectroscopic device 40 for adjusting the excitation wavelength of the excitation light split by the spectroscopic device 40, and a predetermined excitation wavelength spectrally separated by the spectroscopic device 40 And an excitation light irradiation unit 46 for irradiating the measurement target flour MF with the excitation light guided by the optical device 44.

Here, the light source 38 of the spectral illumination unit 22 may be anything as long as it can emit excitation light in a predetermined excitation wavelength range, for example, 200 nm to 900 nm. As the light source 38, for example, a xenon lamp, a tungsten lamp, a wavelength tunable laser, a mercury lamp, a deuterium lamp, or the like can be used.
The spectroscopic device 40 may be any device as long as it can split the excitation light emitted from the light source 38 at a predetermined wavelength interval, for example, an interval of 10 nm within a predetermined excitation wavelength range. As the spectroscopic device 40, for example, an AOTF (Acousto Optic Tunable Filter), a tunable filter (liquid crystal tunable filter), an interference filter, a diffraction grating, or the like can be used.

The excitation wavelength adjusting unit 42 of the spectroscopic device 40 is for arbitrarily setting the excitation wavelength of the excitation light split by the spectroscopic device 40 to a desired excitation wavelength. A manual adjustment unit that can arbitrarily set the excitation wavelength of light itself to a desired excitation wavelength may be used, and the measurer may select an initial value of the excitation wavelength, a variable pitch (excitation wavelength interval to be changed), a final setting value, The excitation wavelength may be automatically changed by inputting or specifying a variable number of times.
The optical device 44 guides the excitation light having a predetermined excitation wavelength dispersed by the spectroscopic device 40 to the excitation light irradiation unit 46, and any optical device 44 can guide the excitation light used for the spectroscopic measurement. However, for example, an optical system including an optical fiber, a lens, a reflecting mirror, and the like can be given.
The excitation light irradiation unit 46 determines the irradiation position of the excitation light so that the measurement target flour MF positioned and supported at the predetermined measurement position of the support unit 21 is appropriately irradiated with the excitation light.
Note that the optical device 44 and the excitation light irradiation unit 46 are not provided when the measurement target flour MF directly supported by the support unit 21 can be appropriately irradiated with the excitation light having a predetermined wavelength separated by the spectroscopic device 40. May be.

The excitation light emitted from the light source 38 is split into excitation light having a predetermined excitation wavelength set by a measurer or automatically by the excitation wavelength adjusting unit 42 in the spectroscopic device 40.
The excitation light converted into the excitation light having a predetermined wavelength in the spectroscopic device 40 is guided by the optical device 44 and is irradiated from the excitation light irradiation unit 46 to the measurement target flour MF. By irradiating with excitation light having a predetermined excitation wavelength, the measurement target flour MF emits fluorescence (having fluorescence intensity) having a pattern peculiar to its component (chemical component) and particle size.
Note that when a wavelength tunable laser is used as the light source 38, excitation light having a desired wavelength is directly obtained from the light source 38, so that the spectroscopic device 40 is unnecessary.
Note that the measurer uses the spectral device 40 having the excitation wavelength adjusting unit 42 with respect to the light source 38, or uses a wavelength tunable laser as the light source 38, so that the wavelength of the excitation light irradiated to the measurement target flour MF. Can be set to any excitation wavelength.
As the spectral illumination unit 22, the spectral illumination device disclosed in Patent Document 1 described above can be used.

  The fluorescence measurement unit 24 of the measurement unit 12 selectively captures a specific fluorescence wavelength among the fluorescence emitted by the measurement target flour MF supported at a predetermined measurement position of the support unit 21 at a predetermined fluorescence wavelength, The fluorescence intensity is measured, and the measurement result is transmitted to the flour discrimination unit 14. In the present invention, in particular, in the stepwise change of the excitation wavelength of the excitation light irradiated to the measurement target flour MF by the spectral illumination unit 22. Accordingly, the fluorescence intensity of the fluorescence generated from the measurement target flour MF irradiated with the excitation light of the predetermined wavelength is measured while changing the fluorescence wavelength to be measured in the predetermined fluorescence wavelength range stepwise. That is, the fluorescence measuring unit 24 can arbitrarily change the fluorescence wavelength for measuring the fluorescence generated from the measurement target flour MF.

As shown in FIG. 4, the fluorescence measurement unit 24 arbitrarily sets the fluorescence wavelength when measuring the fluorescence generated from the measurement target flour MF positioned at the measurement position of the support unit 21 in a predetermined excitation wavelength range. A spectroscopic device 48 that functions as a fluorescent wavelength variable means for changing, and a fluorescence measuring device 50 that measures the fluorescence intensity of the fluorescence having a predetermined fluorescence wavelength generated from the measurement target flour MF and dispersed by the spectroscopic device 48 are provided.
In addition, when measuring the fluorescence intensity of the fluorescence generated from the measurement target flour MF supported at the predetermined measurement position of the support portion 21, depending on the measurement target size of the measurement target flour MF, for example, the size of the flour to be measured is micro In the case of a large size, an optical device may be interposed between the measurement target flour MF and the spectroscopic device 48, and the measurement target size may be expanded to a size suitable for measurement. If the size is macro, it may be reduced. As such an optical device, for example, a microscope, a zoom lens, a macro lens, a wide-angle lens, or the like can be used.

  Here, the spectroscopic device 48 separates the fluorescence generated from the measurement target flour MF irradiated with the excitation light having the predetermined excitation wavelength emitted from the spectral illumination unit 22 into the fluorescence having the predetermined wavelength. The spectroscopic device 48 may be any device as long as it can split the fluorescence generated from the measurement target flour MF at a predetermined wavelength interval, for example, a 10 nm interval within a predetermined fluorescence wavelength range. The spectroscopic device 48 only needs to have means for arbitrarily setting the fluorescence wavelength of the fluorescence to be measured. For example, similar to the spectroscopic device 40, an AOTF (Acousto Optic Tunable Filter), a liquid crystal tunable filter, and an interference filter are used. A diffraction grating or the like can be used.

Moreover, the fluorescence measuring device 50 measures the fluorescence intensity of only the fluorescence having a predetermined wavelength that is selectively dispersed and captured by the spectroscopic device 48 from the fluorescence emitted from the measurement target flour MF.
The fluorescence measuring device 50 may be any device as long as it can measure the fluorescence intensity of fluorescence having a predetermined wavelength. For example, a light detection element such as a photodiode, a phototransistor, or a phototube, or a light sensor such as a CCD sensor or a MOS sensor. In addition, a photomultiplier (photomultiplier tube), a photodetector, or the like can be used. In the case of multipoint measurement, a CCD camera, a scanning image camera, or the like can also be used. In particular, since the photodetector is a single point measurement, the measurement can be performed quickly, and the sensitivity can be increased, so that a slight difference can be discriminated. There are features.

In the fluorescence measurement device 50, only the fluorescence having a predetermined wavelength is selectively captured from the fluorescence emitted from the measurement target flour MF by passing through the spectroscopic device 48, and the fluorescence intensity is measured.
From the fluorescence intensity measured by the fluorescence measuring device 50, information on the fluorescence intensity _{IEx, Em at} a specific excitation wavelength _λEx (irradiation wavelength) and a specific fluorescence wavelength _λEm (measurement wavelength) is obtained. Therefore, by using the fluorescence measuring device 50, the fluorescence intensity at the specific excitation wavelength and fluorescence wavelength of the measurement target flour MF can be measured.
Information (acquired fluorescence intensity) on the measurement target flour MF measured by the fluorescence measurement device 50 is transmitted to the flour discrimination unit 14 that acquires EEM information, and the information acquisition unit of the flour discrimination unit 14 via the control unit 34. 26 and stored directly in the memory 32 as required.

  When acquiring the m-dimensional EEM information at the measurement point of the measurement target flour MF by the fluorescence measurement device 50, the fluorescence intensity of m types of fluorescence is measured while changing the combination of the excitation wavelength and the fluorescence wavelength. In this way, information on the fluorescence intensity of the m types of fluorescence measured by the fluorescence measuring device 50 is transmitted to the flour discrimination unit 14.

By the way, in the spectral illumination unit 22 and the fluorescence measurement unit 24 of the measurement unit 12, the measurement person can arbitrarily change the excitation wavelength to be irradiated and the fluorescence wavelength to be measured manually or automatically. In the present invention, the spectral illumination unit 22 and the fluorescence measurement unit 24 of the measurement unit 12 preferably have means for automatically changing the excitation wavelength and the fluorescence wavelength, respectively. For example, it is preferable that the spectral illumination unit 22 and the fluorescence measurement unit 24 of the measurement unit 12 automatically change the excitation wavelength and the fluorescence wavelength, respectively, based on the control of the control unit 34 of the flour discrimination unit 14.
FIG. 5 is an example of a flowchart of a process of measuring the fluorescence intensity while automatically changing the excitation wavelength and the fluorescence wavelength in the measurement unit 12 under the control of the control unit (see the above-described fluorescence intensity measurement process of the present invention). It is.

First, in order to automatically change the excitation wavelength of the spectroscopic illumination unit 22 and the fluorescence wavelength of the fluorescence measurement unit 24, the measurer uses the operation unit 16 such as a keyboard / mouse to measure the excitation wavelength range and excitation wavelength of the EEM to be measured. A pitch is input (step S50). Next, the fluorescence wavelength range and the fluorescence wavelength pitch are input (step S52).
Next, the control unit 34 sets the minimum excitation wavelength and the minimum fluorescence wavelength to initial values for the spectral illumination unit 22 and the fluorescence measurement unit 24 based on the input excitation wavelength range, fluorescence wavelength range, and each wavelength pitch. (Step S54), and instructs to measure the first fluorescence intensity (Step S56).

  If the fluorescence wavelength does not fall outside the set fluorescence wavelength range even if the fluorescence wavelength is increased by one pitch (N in step S58), the control unit 34 increases the fluorescence wavelength by one pitch to the fluorescence measurement unit 24. An instruction is given to measure the second fluorescence intensity (steps S60 and S56). Similarly, when the fluorescence wavelength is increased by one pitch, the measurement of the fluorescence intensity is repeated by increasing the fluorescence wavelength by one pitch until the fluorescence wavelength is out of the set fluorescence wavelength range (N in step S58, step S60, step S56).

When the fluorescence wavelength is increased by one pitch, the fluorescence wavelength is outside the set fluorescence wavelength range (Y in step S58), and even if the excitation wavelength is increased by one pitch, the wavelength is not outside the set excitation wavelength range. In the case (N in step S62), the control unit 34 increases the excitation wavelength by one pitch in the spectral illumination unit 22, and sets the fluorescence wavelength to the initial value (here, the initial value of the fluorescence wavelength is the fluorescence wavelength). (It means the same wavelength as the excitation wavelength increased by one pitch) (step S64), and commands the fluorescence measurement unit 24 to measure the fluorescence intensity (step S56). Thereafter, when the excitation wavelength is increased by 1 pitch, the measurement is repeated by increasing the excitation wavelength by 1 pitch until the wavelength is outside the set excitation wavelength range (N in step S62, step S64).
When the fluorescence wavelength is increased by one pitch, the fluorescence wavelength is outside the set fluorescence wavelength range (Y in step S58), and when the excitation wavelength is increased by one pitch, the wavelength is outside the set excitation wavelength range. In the case (Y of step S62), the measurement is terminated, and the measured value of the measured fluorescence intensity is transmitted to the flour discrimination unit 14.

In the present embodiment, the spectroscopic illumination unit 22 that generates fluorescence from the flour (each component) to be measured by irradiating excitation light of a predetermined wavelength and the fluorescence emitted from the flour to be measured are predetermined. Although the apparatus has the fluorescence measuring unit 24 that selectively captures fluorescence having a wavelength of 5 and measures the fluorescence intensity, the apparatus may be integrated into one apparatus.
Moreover, the measurement unit 12 mentioned above implements the reference fluorescence characteristic acquisition and storage process (refer FIG. 1) of step S10, and parameterizes according to the particle size of the known flour used as the object of collation, and the kind of kind beforehand. It is also possible to use the obtained fluorescence characteristics as reference fluorescence characteristics.
The measurement unit 12 used in the present invention is basically configured as described above.

Next, each component of the flour discrimination unit 14 will be described.
The information acquisition unit 26 acquires EEM information of the measurement target flour MF from the measured values of the fluorescence intensity I _{Ex, Em at} a specific excitation wavelength λ _Ex (irradiation wavelength) and a specific fluorescence wavelength λ _Em (measurement wavelength). Then, it is for implementing the post-process after the fluorescence intensity measurement of the EEM information acquisition process (refer FIG. 1) of step S14 of this invention mentioned above. That is, the information acquisition unit 26 is obtained by the fluorescence measurement unit 24 of the measurement unit 12 while gradually changing the excitation wavelength and the fluorescence wavelength in the predetermined excitation wavelength range and the predetermined fluorescence wavelength range in the measurement unit 12. The measured fluorescence intensity is received from the fluorescence measuring unit 24, and EEM information of the measurement target flour MF is acquired from the received measured fluorescence intensity.

The fluorescence characteristic extraction unit 28 performs statistical analysis processing on the EEM information of the measurement target flour MF acquired by the information acquisition unit 26 and extracts the fluorescence characteristic of the measurement target flour MF as a parameter. This is for carrying out the fluorescence characteristic extraction step (see FIG. 1) in step S16.
Here, the fluorescence characteristic extraction unit 28 preferably acquires the fluorescence characteristic of the flour MF to be measured from the EEM information as a two-dimensional parameter, but may acquire it as a multidimensional parameter if the characteristic can be accurately grasped.

In addition, the fluorescence characteristic extraction unit 28 preferably performs statistical analysis processing by multivariate analysis or data mining as fluorescence characteristic extraction. Data structure analysis, discriminant analysis, pattern classification, multi-factor data analysis, regression analysis, and learning machine It is preferable to use at least one method selected from the group consisting of:
Furthermore, the fluorescence characteristic extraction unit 28 preferably performs the data structure analysis using at least one method selected from the group consisting of principal component analysis, factor analysis, correspondence analysis, and independent component analysis.
Furthermore, the fluorescence characteristic extraction unit 28 performs statistical analysis processing using the principal component analysis technique as fluorescence characteristic extraction, calculates two principal component scores from the EEM information as two-dimensional parameters, and calculates the calculated two It is preferable to acquire the main component score as the fluorescence characteristic of the measurement target flour MF.

Furthermore, the fluorescence characteristic extraction unit 28 preferably performs discriminant analysis using a linear discriminant analysis method or a nonlinear discriminant analysis method.
Furthermore, the fluorescence characteristic extraction unit 28 performs multivariate analysis or the data mining as a fluorescence characteristic extraction using a discriminant analysis technique, and EEM information is preliminarily stored in the memory (DB) 32 and / or the external memory (DB). ) By calculating the parameters by applying to the discriminant stored in 36 and comparing with the parameters stored in advance in the memory (DB) 32 and / or the external memory (DB) 36, the type of the flour is discriminated. Is preferred.
Furthermore, the fluorescence characteristic extraction unit 28 preferably performs pattern classification by cluster analysis or multidimensional scaling.
Furthermore, the fluorescence characteristic extraction unit 28 preferably performs the regression analysis using a linear regression, nonlinear regression, or multiple regression analysis technique.
Furthermore, the fluorescence characteristic extraction unit 28 performs multivariate analysis or the data mining as a fluorescence characteristic extraction by using a method of multiple regression analysis, and EEM information as a parameter is stored in advance in the database. It is preferable to calculate by fitting to a line and collate with a parameter stored in advance in the memory (DB) 32 and / or the external memory (DB) 36.

Furthermore, it is preferable that the fluorescence characteristic extraction unit 28 performs the learning machine using at least one method selected from the group consisting of a neural network, a self-organizing map, group learning, and a genetic algorithm.
In addition, the information acquisition part 26 and the fluorescence characteristic extraction part 28 of the flour discrimination | determination unit 14 mentioned above implement the reference fluorescence characteristic acquisition and storage process (refer FIG. 1) of step S10, and are known beforehand used as the object of collation. It can also be used when parameterized fluorescence characteristics corresponding to the grain size of the flour and the type of varieties are calculated as reference fluorescence characteristics.

The fluorescence characteristic collation unit 30 uses the particle size obtained by the particle size acquisition unit 20 to change the parameterized fluorescence characteristic of the measurement target flour MF extracted by the fluorescence characteristic extraction unit 28 according to the type of flour and the particle size thereof. The type of varieties of the measurement target flour MF is discriminated by comparing with the previously obtained parameterized fluorescence characteristics (reference fluorescence characteristics) of known flour, and the fluorescence of step S18 of the present invention described above is determined. This is for carrying out a characteristic matching step (see FIG. 1).
In addition, the fluorescence characteristic collation part 30 makes the fluorescence characteristic of the measurement object flour MF extracted by the fluorescence characteristic extraction part 28 based on the flour for each component of the grain and the grain size of the known grain that has been obtained in advance. It is also possible to identify the flour to be measured MF as a known flour containing as a main component a specific component of a known grain by collating with the parameterized fluorescence characteristics of the known flour containing each component as a main component. .

  In addition, as above-mentioned, collation with the fluorescence characteristic of the measurement object flour MF in the fluorescence characteristic collation part 30 and the reference fluorescence characteristic of known flour uses the particle size and fluorescence characteristic (parameter) of the measurement object flour MF as search data. The memory (DB) 32 and / or the external memory (DB) 36 in which the grain sizes and reference fluorescence characteristics of many known varieties are stored are searched and matched, or the matching degree is equal to or greater than a predetermined threshold The type of flour variety may be specified by searching for the type of flour, and the type of the predetermined variety of flour may vary depending on the particle size of the flour. In the case of being given as a plane area (two-dimensional parameter), what is included in the space area or the plane area may be obtained to specify the type of flour variety.

The control unit 34 includes a particle size acquisition unit 20, a spectral illumination unit 22 and a fluorescence measurement unit 24 of the measurement unit 12, and an information acquisition unit 26, a fluorescence characteristic extraction unit 28, a fluorescence characteristic verification unit 30 and a flour discrimination unit 14. While controlling each part of the memory (DB) 32 and further the external memory (DB) 36, the whole flour discriminating device 10 of the present invention is also controlled, and the discriminating device 10 carries out the flour discriminating method of the present invention. Is to control.
That is, the controller 34 operates the instruction to the particle size acquisition unit 20 to acquire the particle size of the measurement target flour MF, or the spectroscopic illumination unit 22 and the fluorescence measurement unit 24 of the measurement unit 12. Using the excitation wavelength range, fluorescence wavelength range, and wavelength pitch input by the unit 16, the excitation wavelength to be measured and the fluorescence wavelength to be measured are adjusted to measure the fluorescence data of the measurement target flour MF, and further, information acquisition An instruction or command for acquiring EEM information from the measurement value of the fluorescence intensity transmitted from the fluorescence measurement unit 24 can be issued to the unit 26.

The control unit 34 further pre-processes the measurement value of the fluorescence intensity transmitted from the fluorescence measurement unit 24 to the information acquisition unit 26, and deletes the influence of disturbance light from the measurement value of the fluorescence intensity. Or an instruction or command for performing a statistical analysis process on the EEM information acquired from the fluorescence intensity measurement result or the pre-processed EEM information.
In addition, the control unit 34 issues an instruction or command to perform statistical analysis processing for parameterizing the EEM information obtained by the information acquisition unit 26 to the fluorescence characteristic extraction unit 28, and also sends the instruction to the fluorescence characteristic verification unit 30. On the other hand, the parameterized fluorescence characteristics are compared with the reference fluorescence characteristics of known varieties obtained in advance and the flour of the particle size, and instructions and commands for determining the type of the flour MF to be measured are issued. Can do.

The memory 32 measures the particle size of the measurement target flour MF acquired by the particle size acquisition unit 20, the EEM information of the measurement target flour MF acquired by the information acquisition unit 26 of the flour discrimination unit 14, and the fluorescence characteristic extraction unit 28. Database for storing parameterized fluorescence characteristics of target flour MF, and database for storing types, grain sizes, reference fluorescence characteristics, etc. ) As a function.
That is, the memory 32 stores the value of the particle size transmitted from the particle size acquisition unit 20 and the measurement result (measurement value) of the fluorescence intensity measured by the fluorescence measurement unit 24 of the measurement unit 12 and transmitted from the fluorescence measurement unit 24. The EEM information obtained in the information acquisition unit 26 from the measurement result of the fluorescence intensity transmitted from the fluorescence measurement unit 24 can also be stored.

In addition, the memory 32 has a parameterization fluorescence characteristic of the measurement target flour MF extracted by performing statistical analysis processing on the EEM information of the flour acquired by the information acquisition unit 26 in the fluorescence characteristic extraction unit 28, and a fluorescence characteristic matching unit 30. It is also possible to store the type, particle size, and fluorescence characteristics of the measurement target flour MF identified by collating in (1).
Further, the memory 32 stores, as a database, the types of a variety of known flour varieties obtained as a comparison reference (in the case of flour for each component of a known cereal, the type of each component) and the granularity thereof. Three of the reference fluorescence characteristics (parameterized fluorescence characteristics) of the corresponding known flour (flour containing each component as a main component) can also be stored as a unit.

  Moreover, the known flour that can be used when extracting the fluorescence characteristics of an unknown sample in the fluorescence characteristics extraction step, as well as storing the reference fluorescence characteristics itself according to the type and grain size of the flour varieties. General formula Y = f (x) (x: EEM information, Y: output information) created based on EEM information obtained from the particle size and type of the reference, and the reference fluorescence according to the type and particle size of the flour variety The spatial area in the spatial coordinates of the characteristic and the planar area in the plane coordinates (distribution status of EEM information obtained from known grain size and type of flour, etc.) can be stored (stored) in the memory 32. .

The external memory 36 stores a database of three types of known flour varieties, particle sizes, and reference fluorescence characteristics as a database instead of or in addition to the memory 32. Is. In addition, when storing the types, grain sizes, and reference fluorescence characteristics of many known flour varieties to be compared in the external memory 36, it is not necessary to store these data in the memory 32. Since the function need not be provided, the capacity of the memory 32 can be reduced. Therefore, for the above reasons, the external memory 36 may not be provided, but is preferably provided.
In addition, the external memory 36 may store the measurement value of the particle size, the measurement result of the fluorescence intensity, various EEM information, the parameterized fluorescence characteristics, the type of the specified flour MF to be measured, the particle size, the fluorescence characteristics, and the like. good.
The flour discrimination unit 14 is basically configured as described above.

The operation unit 16 includes a keyboard, a mouse, and the like. The measurer gives instructions to the determination device 10 such as particle size acquisition by the particle size acquisition unit 20, measurement operation by the measurement unit 12, and determination operation by the flour determination unit 14. Input, input for various instructions and instructions for implementing the method for discriminating flour of the present invention, and the types, particle sizes, and reference fluorescence characteristics of many known flour varieties for database creation, etc. This is for performing the input.
Further, the display 18 displays the value of the obtained particle size by the particle size obtaining unit 20 and the measurement result by the measurement unit 12, for example, the measurement result of the fluorescence intensity by the fluorescence measurement unit 24, and various results by the flour discrimination unit 14, for example, , Various EEM information by the information acquisition unit 26, parameterized fluorescence characteristics by the fluorescence characteristic extraction unit 28, types of the measurement target flour MF identified by the fluorescence characteristic collation unit 30, grain size, fluorescence characteristics, etc. 16 is for displaying part or all of the operation instructions and inputs by 16. Note that the display 18 may be configured to display not only the input result from the keyboard or mouse of the operation unit 16 but also a GUI that supports the input by the operation unit 16.

Below, an example of the discrimination operation of the flour by the flour discrimination unit 14 is demonstrated.
In the present invention, when the measurement value of the fluorescence intensity of the measurement target flour MF is obtained in the measurement unit 12, the measurement result of the fluorescence intensity transmitted from the fluorescence measurement unit 24 of the measurement unit 12 to the flour discrimination unit 14 is stored in the memory. 32.
Here, when an instruction or command for determining the flour by the flour determining unit 14 is issued from the operation unit 16 to the control unit 34 by the measurer, the control unit 34 is stored in the memory 32 in the information acquisition unit 26. The measurement result of fluorescence intensity is read, and an instruction is given to acquire m-dimensional EEM information. As described above, the information acquisition unit 26 acquires m-dimensional EEM information from the read measurement result of fluorescence intensity, and stores the acquired EEM information in the memory 32.

Next, when the control unit 34 instructs the fluorescence characteristic extraction unit 28 to extract the fluorescence characteristic of the m-dimensional EEM information, the fluorescence characteristic extraction unit 28 stores the m-dimensional EEM information stored in the memory 32. The fluorescence characteristic is acquired as a parameter, for example, a two-dimensional parameter, by a statistical analysis process such as a principal component analysis method that has been read and set by default or specified by the measurer, and the acquired parameterized fluorescence characteristic is stored in the memory 32 To do.
Further, the fluorescence characteristic extraction unit 28 performs statistical analysis processing on the EEM information obtained by the information acquisition unit 26 to extract the parameterized, preferably two-dimensional parameterized fluorescence characteristics, and transmits them to the display 18. Can be displayed above.

Next, when the control unit 34 instructs the fluorescence characteristic matching unit 30 to determine the type of flour of the measurement target flour MF, the fluorescence characteristic matching unit 30 includes the parameterized fluorescence characteristics stored in the memory 32, The reference fluorescence characteristics (parameters) of the same or substantially the same grain size of known flour stored as a database in the memory 32 or the external memory 36 are read one after another, and these are used to calculate the corresponding known flour for each statistical processing method. If the type of cultivar is specified as the type of measurement target flour MF, or if there is no matching known flour reference fluorescence characteristic, the type of measurement target flour MF is determined to be indistinguishable or indistinguishable. End the operation.
In addition, the collation method by each statistical analysis processing method is explained in full detail in an Example.

  Examples of the present invention will be specifically described below, but the present invention is not limited to these examples.

[Example 1]
Using the flour product as the measurement target flour MF, the measurement target flour MF was determined.
As an unknown sample, flour products were identified using flour for pasta (durum flour, trade name: DF, manufactured by Nisshin Flour Mills) and bread flour (strong flour: trade name, Camellia, manufactured by Nisshin Flour Mills).
The average particle sizes of the pasta flour and bread flour used were 68 μm and 65 μm, respectively.
First, a pasta flour sample and a bread flour sample were each enclosed in a quartz cell. A quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12 of the discrimination device 10 shown in FIG. 1, and the spectral illumination unit of the measurement unit 12 is placed on each sample. 22 and the fluorescence measurement part 24 were used, and the fluorescence intensity measurement of the EEM information acquisition process of step 14 of the flour discrimination | determination method of this invention mentioned above was implemented on the following measurement conditions, and the fluorescence intensity was measured 2 times.
Measurement conditions Excitation wavelength range: 200 to 900 nm / 10 nm interval Fluorescence wavelength range: 200 to 900 nm / 10 nm interval Scan speed: 1000 nm / min

Here, a spectroscopic measurement device in which the spectroscopic illumination unit 22, the measurement target flour support unit 21 and the fluorescence measurement unit 24 of the measurement unit 12 are integrated is used.
Further, as the flour discrimination unit 14 of the discrimination device 10 shown in FIG. 3, a commercially available personal computer in which flour discrimination software developed by the present inventors was incorporated was used.

EEM information of each material was obtained from the measured value (measurement result) of the fluorescence intensity of each sample thus obtained. The results are shown in FIGS. 6 (a) and 6 (b). 6 (a) and 6 (b) show the results of a pasta flour sample and a bread flour sample, respectively.
From this EEM information, principal component analysis was performed as statistical analysis processing, and two-dimensional parameters were obtained as parameterized fluorescence characteristics. The result is shown in FIG.

In addition, as a reference flour (known sample) for known flour products, a plurality of known pasta flours and a plurality of known bread flours are used, and each reference is made with the same apparatus configuration and the same measurement conditions. The measurement result of the fluorescence intensity of flour was obtained, and EEM information of each reference flour was obtained. Here, as the reference flour of pasta flour and bread flour measured for reference, those having the same average particle size as the measurement target flour were used. The obtained EEM information of each reference flour was subjected to the same principal component analysis as statistical analysis processing, and the reference fluorescence characteristics (two-dimensional parameters) for each flour product were obtained. The result is shown in FIG.
By comparing FIGS. 7 and 8, the fluorescence characteristics of each sample (two-dimensional coordinates represented by two-dimensional parameters (principal component axis 1 (contribution rate 96.6%) and principal component axis 2 (contribution rate 3. 3%))))) is in a 95% probability ellipse (range in which 95% of the reference fluorescence characteristic of the known sample is distributed) within the area of the reference fluorescence characteristic of each flour product (known sample) obtained in advance Since it was plotted, it was confirmed that pasta flour and bread flour could be identified and identified.
A method for collating unknown samples with known samples by principal component analysis will be described in detail in Example 4.

[Example 2]
As the measurement target flour MF, the measurement target flour MF was determined using flour having different particle sizes.
As an unknown sample, dried noodles (trade name No. 1 manufactured by Nisshin Foods) were first coarsely pulverized to prepare a coarse particle powder (average particle size 200 μm), and then further pulverized to obtain a fine particle particle (average) What adjusted the particle size of 80 micrometers was used.
First, a coarse powder sample and a fine powder sample were each enclosed in a quartz cell.
Using the same discriminating apparatus 10 as in the first embodiment, the quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12, and the first embodiment is applied to each sample. Fluorescence measurement was carried out under the same measurement conditions as above, and the fluorescence intensity of these powder samples was measured. From the results, EEM information was obtained in the same manner as in Example 1. The results are shown in FIGS. 9 (a) and 9 (b). 9A and 9B show the results of a coarse powder sample and a fine powder sample, respectively.

Further, from the EEM information obtained from the fluorescence intensity measurement results, a principal component analysis was performed in the same manner as in Example 1 to obtain two-dimensional parameters as parameterized fluorescence characteristics. The result is shown in FIG. As a known sample, a dry noodle powder sample having the same particle size as that of the unknown sample (average particle size of 80 μm or average particle size of 200 μm) is prepared separately, and in the same manner as in Example 1, for each dry noodle powder sample having a different particle size, Reference fluorescence characteristics (two-dimensional parameters) of a plurality of dry noodle powder samples were obtained. The result is shown in FIG.
By comparing FIGS. 10 and 11, the fluorescence characteristics of each sample (two-dimensional coordinates represented by two-dimensional parameters (principal component axis 1 (contribution rate 86.7%) and principal component axis 3 (contribution rate 2. 5%)))) is a 95% probability ellipse in the region of the reference fluorescence characteristic of the powder of each particle size (known sample) obtained in advance (a range in which 95% of the reference fluorescence characteristic of the known sample is distributed) Therefore, it was confirmed that coarse powder and fine powder could be distinguished and identified.
A method for collating unknown samples with known samples by principal component analysis will be described in detail in Example 4.

Example 3
Differentiated flour samples (flours of different purity, that is, a mixture of pure flour and heterogeneous flour (also referred to as allergen-containing flour)) were used to determine flour.
As an unknown sample, a sample of 100% of domestic buckwheat flour (trade name: Shirasina flour Ishimori Flour Mills) and a mixed sample of 30% of wheat flour (trade name: Camellia Nisshin Flour Mills) mixed with the same buckwheat flour Two samples were prepared. In addition, the average particle diameters of the used oat flour and wheat flour were 90 μm and 63 μm, respectively.
First, a pure flour sample made from 100% domestic oat flour and a mixed flour sample containing 30% wheat flour in oat flour were each enclosed in a quartz cell.
Using the same discriminating apparatus 10 as in the first embodiment, the quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12, and the first embodiment is applied to each sample. Fluorescence measurement was performed under the same measurement conditions as above, and the fluorescence intensity of these flour samples was measured. From the results, EEM information was obtained in the same manner as in Example 1. The results are shown in FIGS. 12 (a) and 12 (b). In addition, Fig.12 (a) and (b) show the result of the pure flour sample only of buckwheat flour, and a mixed flour sample, respectively.

Further, from the EEM information obtained from the fluorescence intensity measurement results, a principal component analysis was performed in the same manner as in Example 1 to obtain two-dimensional parameters as parameterized fluorescence characteristics. The result is shown in FIG.
In addition, as a known sample, a pure flour sample only of buckwheat flour and a similar mixed flour sample are separately prepared in advance, and in the same manner as in Example 1, a plurality of flour samples are obtained for each flour sample of various particle sizes. Reference fluorescence characteristics (two-dimensional parameters) were determined. Among them, FIG. 14 shows the results of the reference fluorescence characteristics (two-dimensional parameters) when the average particle sizes of the reference flours of oat flour and wheat flour are 90 μm and 63 μm, respectively.
By comparing FIGS. 13 and 14, the fluorescence characteristics of each flour sample (two-dimensional coordinates represented by two-dimensional parameters (principal component axis 1 (contribution rate 71.3%)) and principal component axis 2 (contribution rate 10 0.0%)))) is a 95% probability ellipse within the area of the reference fluorescence characteristic of each flour sample (known sample) obtained in advance (range in which 95% of the reference fluorescence characteristic of the known sample is distributed) Therefore, it was confirmed that a mixed flour sample of 100% buckwheat flour and a mixed flour sample of buckwheat flour containing 30% wheat flour can be identified and specified.
A method for collating unknown samples with known samples by principal component analysis will be described in detail in Example 4.

Example 4
EEM information was obtained by the same method as in Example 1, and the different types of flour were discriminated using principal component analysis as a statistical analysis processing method.
As an unknown sample, buckwheat flour (trade name, manufactured by Ishimori Flour Milling Co., Ltd.), rye flour (trade name, mail Dunkel, manufactured by Nisshin Flour Milling Co., Ltd.), strong flour (trade name, manufactured by Camellia Nisshin Flour Milling Co., Ltd.) and rice flour (trade name, Refrine) Gunma Flour Milling) was prepared. In addition, the average particle diameter of each flour is 90 micrometers, 90 micrometers, 63 micrometers, and 48 micrometers, respectively.
First, each sample was enclosed in a quartz cell.
Using the same discriminating apparatus 10 as in the first embodiment, the quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12, and the first embodiment is applied to each sample. Fluorescence measurement was performed under the same measurement conditions as above, and the fluorescence intensity of these flour samples was measured. From the results, EEM information was obtained in the same manner as in Example 1.
In the same way, EEM information of a known sample is acquired in advance by the same method, and a mathematical expression (eigenvector) that is converted to a principal component coordinate system based on the EEM information: x is EEM information, and Y is an output result. Y = f (x) is created, and the distribution state of each sample group in the principal component coordinate system is stored in the memory (DB) 32 and / or the external memory (DB) 36. As the known sample, two samples each of the above-described samples were used. FIG. 15 shows the distribution of each known sample group in the principal component coordinate system when the average particle sizes of oat flour, rye flour, strong flour and rice flour measured for reference are 90 μm, 90 μm, 63 μm and 48 μm, respectively. Show.

When the output result obtained by substituting EEM information of an unknown sample into the above general formula is plotted on the principal component coordinate system created based on the known sample shown in FIG. The type of flour is determined by whether or not the position in the condition satisfies the following conditions (1) to (3).
(1) Each principal component score of an unknown sample falls within the mean ± 2 × standard deviation.
(2) The plot of the unknown sample exists inside the 95% established ellipse (the range in which 95% of the known sample is distributed) in the plot diagram of the known sample.
(3) Each principal component score of the unknown material is closest to the center of gravity of each principal component score of the known sample.
In addition, the range as described in said (1) and (2) can be set arbitrarily. (For example, mean ± 2.5 × standard deviation, 90% probability ellipse)
In this example, as shown in FIG. 15, it was confirmed that each unknown sample belongs to each known sample.

Example 5
Using cluster analysis as a statistical analysis processing method, the attributes of the unknown sample (flour) were determined. Cluster analysis is an analysis method in which samples that are close to each other in a multidimensional space are connected one after another to form one cluster.
As unknown samples, buckwheat flour (trade name, manufactured by Ishimori Flour Milling Co., Ltd.), rye flour (trade name, manufactured by Meer Dunkel Nisshin Flour Milling Co., Ltd.) and rice flour (trade name, manufactured by Referrine Gunma Flour Milling Co., Ltd.) were prepared.
In addition, the average particle diameter of each flour is 90 micrometers, 90 micrometers, and 48 micrometers, respectively.
In addition, EEM information of each known sample (buckwheat flour, rye flour and rice flour) is acquired in the same manner as in Example 1, and the distribution of the known samples classified into each cluster based on the EEM information. The situation (average value, standard deviation, etc.) is stored in the memory (DB) 32 and / or the external memory (DB) 36. The known sample used was the same sample and particle size as the unknown sample. FIG. 16 (a) shows the classification status of each cluster when the average particle diameters of buckwheat flour, rye flour, and rice flour as reference flours are 90 μm, 90 μm, and 48 μm, respectively.
Based on the clusters of known samples stored in the database, the attributes of each unknown sample are determined according to which cluster the three unknown samples are combined.
In this example, as shown in FIG. 16B, when three pieces of verification data were added, it was confirmed that each sample was bound to each cluster.

Example 6
EEM information was obtained by the same method as in Example 1, and different types of flour were discriminated using discriminant analysis as a statistical analysis processing method. Discriminant analysis is a method of creating a discriminant for discriminating each group of known samples using a known sample and applying the EEM information of the unknown sample to the discriminant to discriminate the sample group.
As an unknown sample, buckwheat flour (trade name, manufactured by Ishimori Flour Milling Co., Ltd.), rye flour (trade name, mail Dunkel, manufactured by Nisshin Flour Milling Co., Ltd.), strong flour (trade name, manufactured by Camellia Nisshin Flour Milling Co., Ltd.) and rice flour (trade name, Refrine) Gunma Flour Milling) was prepared. In addition, the average particle diameter of each flour is 90 micrometers, 90 micrometers, 63 micrometers, and 48 micrometers, respectively.
First, each sample was enclosed in a quartz cell.
Using the same discriminating apparatus 10 as in the first embodiment, the quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12, and the first embodiment is applied to each sample. Fluorescence measurement was performed under the same measurement conditions as above, and the fluorescence intensity of these flour samples was measured. From the results, EEM information was obtained in the same manner as in Example 1.
In addition, by the same method, EEM information of each known sample (buckwheat flour, rye flour, strong flour and rice flour) is acquired for every two samples in advance, and x is EEM information and Y is output based on the EEM information. The resulting general formula (discriminant) Y = f (x) is created, and the distribution state of each sample group in the discriminant analysis is stored in the memory (DB) 32 and / or the external memory (DB) 36. . The known sample used was the same as the sample whose particle size was unknown. FIG. 17 shows the distribution state of each group in the discriminant analysis when the average particle sizes of the known samples of buckwheat flour, rye flour, strong flour and rice flour are 90 μm, 90 μm, 63 μm and 48 μm, respectively.

  The output result obtained by substituting the EEM information of each flour that is an unknown sample into the above general formula (discriminant) is shown as the probability that the unknown sample belongs to each known sample group. Accordingly, when the output result of the unknown sample is superimposed on the distribution state of each sample group of the discriminant analysis created based on the known sample shown in FIG. 17, the sample group having the largest establishment (output result) is plotted. Since the unknown sample belongs, the type of flour can be determined. FIG. 18 shows the discrimination result of each unknown sample. In this example, as shown in FIG. 18, it was confirmed that each unknown sample was discriminated with 100% accuracy.

Example 7
EEM information was obtained by the same method as in Example 1, and different types of flour were discriminated using regression analysis (PLS regression analysis) as a statistical analysis processing method. Regression analysis creates a calibration curve that predicts the amount of measurement object (mixing ratio of the measurement object) from EEM information of a known sample, and applies the EEM information of an unknown sample to the calibration curve, and the content of the measurement object This is a method for discriminating.
As an unknown sample, a mixed sample in which buckwheat flour (trade name, manufactured by Ishimori Flour Milling Co., Ltd.) and strong powder (product name: Camellia Nisshin Flour Milling Co., Ltd.) was prepared. In addition, the average particle diameter of each flour is 90 micrometers and 63 micrometers, respectively.
First, the sample was enclosed in a quartz cell.
Using the same discriminating apparatus 10 as in the first embodiment, the quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12, and the first embodiment is applied to each sample. Fluorescence measurement was performed under the same measurement conditions as above, and the fluorescence intensity of these flour samples was measured. From the results, EEM information was obtained in the same manner as in Example 1.

In the same way, the particle size and EEM information of the strong powder (known sample) are acquired in advance, and a calibration curve for predicting the blending ratio of the strong powder as the measurement object is created based on the EEM information. (DB) 32 and / or external memory (DB) 36 is stored. A calibration curve created with a known sample is shown in FIG. The known sample was the same as the strong powder of the unknown sample. In addition, the same particle size as that of the unknown sample was used.
Substituting the EEM information of the strong powder, which is an unknown sample, into a calibration curve, and obtaining the output result, that is, the content (predicted value) of the component contained in the unknown sample, the predicted value as shown in FIG. Obtained. As a result of collating the calibration curve of the known sample with the predicted value of the unknown sample, it was confirmed that the content of strong powder in the unknown sample can be predicted with a determination coefficient of 0.99 or more and an error of 3.59%.

Example 8
EEM information was acquired by the same method as in Example 1, and a heterogeneous flour was discriminated using a neural network as a statistical analysis processing method.
As shown in FIG. 20, the neural network constructs a network structure composed of an input layer, an intermediate layer, and an output layer, and inputs EEM information of strong powder using a known sample, thereby obtaining a response of “strong powder”. This is a method for calculating such a weight coefficient w.
In this example, as an unknown sample, buckwheat flour (trade name, manufactured by Ishimori Flour Milling Co., Ltd.), rye flour (trade name: mail Dunkel, manufactured by Nisshin Flour Milling Co., Ltd.), strong flour (trade name: Camellia Nisshin Flour Milling Co., Ltd.) and Rice flour (trade name: Refrine, Gunma Flour Mills) was prepared.
In addition, each average particle diameter of buckwheat flour, rye flour, strong flour and rice flour is 90 μm, 90 μm, 63 μm and 48 μm, respectively.
First, each sample was enclosed in a quartz cell.
Using the same discriminating apparatus 10 as in the first embodiment, the quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12, and the first embodiment is applied to each sample. Fluorescence measurement was performed under the same measurement conditions as above, and the fluorescence intensity of these flour samples was measured. From the results, EEM information was obtained in the same manner as in Example 1.

In addition, by the same method, the particle size and EEM information of each known sample (buckwheat flour, rye flour, strong flour and rice flour) are acquired in advance, and based on the EEM information, x is EEM information and Y is the output result. The following general formula Y = f (x) is created, and the network structure, the weighting coefficient, and the general formula are stored in the memory (DB) 32 and / or the external memory (DB) 36. The known sample was the same as the unknown sample. In addition, the same particle size as that of the unknown sample was used.
In the output layer of the above network structure, there are four nodes corresponding to each group of buckwheat flour, rye flour, strong flour and rice flour. When EEM information of unknown samples is input to the above general formula, it belongs to each known sample group. Since the probability is obtained as an output result and the unknown sample belongs to the sample group having the largest probability (output result), the type of flour can be predicted. FIG. 21 shows the prediction results for each unknown sample.
In this example, as shown in FIG. 21, it was confirmed that an unknown sample can be correctly identified as each flour with 100% accuracy.

Example 9
EEM information was acquired in the same manner as in Example 1, and flours having different particle sizes were determined using principal component analysis as a statistical processing method.
As the flour of the unknown sample, Durum semolina was pulverized with a pulverizer and divided into fractions with particle sizes corresponding to the following four zones, and used. Each particle size was measured and confirmed by the particle size acquisition unit 20. As the particle size acquisition unit 20, a particle size measurement device (trade name, manufactured by Microtrac Nikkiso Co., Ltd.) was used.
Zone (1): fraction with coarse particle size 300 μm or more (average particle size 500 μm)
Zone (2): fraction with a slightly coarse particle size 150 μm or more and less than 300 μm (average particle size 200 μm)
Zone (3): Fine particle size fraction 10 μm or more and less than 150 μm (average particle size 60 μm)
Zone (4): fraction with very fine particle size <10 μm (average particle size 1 μm)
First, a flour sample of each particle size fraction was enclosed in a quartz cell.
Using the same discriminating apparatus 10 as in the first embodiment, the quartz cell in which each sample is enclosed is positioned and placed at a predetermined measurement position of the support portion 21 of the measurement unit 12, and the first embodiment is applied to each sample. Fluorescence measurement was performed under the same measurement conditions as above, and the fluorescence intensity of these flour samples was measured. From the results, EEM information was obtained in the same manner as in Example 1.

Further, from the EEM information obtained from the fluorescence intensity measurement results, a principal component analysis was performed in the same manner as in Example 1 to obtain two-dimensional parameters as parameterized fluorescence characteristics. The result is shown in FIG. In addition, as a known sample, a flour sample of durum semolina having substantially the same particle size (average particle size) is prepared separately in advance. The fluorescence characteristics (two-dimensional parameters) were obtained.
Fluorescence characteristics of each sample shown in FIG. 22 (position on two-dimensional coordinates represented by two-dimensional parameters (principal component axis 1: contribution ratio 82.3%, principal component axis 2: contribution ratio 4.7%)) Was almost plotted in the region of the reference fluorescence characteristics of flour of each particle size given in advance, it was confirmed that flour from zones (1) to (4) can be discriminated and specified.
The method for collating unknown samples with known samples by principal component analysis is as described in detail in the fourth embodiment.

The flour discriminating method and apparatus of the present invention can be used for the type of flour, for example, the final product variety of flour, or the intermediate product (flour) produced during the production of flour, without cost, labor and time. The type of varieties and varieties containing specific components can be easily and reliably discriminated in real time, so quality control of flour products such as flour and the production process (milling process) of flour products ) Management and control, in particular, not only the final process, but also the management and control of the intermediate production flour production process (milling intermediate process), the quality control of the intermediate production flour, and the like.
The flour discrimination method and apparatus of the present invention can be used for discrimination and identification of powder (flour) of each component of grain, or for discrimination and identification of flour containing allergen and allergen in flour.
In particular, since the present invention prepares in advance the types, particle sizes, and reference fluorescence characteristics of known flour varieties, intermediate flour for discrimination of flour products such as flour products and management of the production process of flour products Discrimination, flour consisting of each component of grain or flour containing each component as a main component, discrimination of flour containing known allergens and other known components, and allergens and other contained in flour according to this discrimination result It can be suitably used for identifying components and the like.

Therefore, the present invention provides a flour, a component in the flour mixture and its components and heterogeneous flour, for example, a flour in rye flour, a simple detection technology for flour in durum flour, a simple detection technology, a flour, and a bean flour in the flour mixture. , Rice flour, corn flour simple discrimination technology and simple detection technology, wheat flour in buckwheat flour, simple discrimination technology and simple detection technology of wheat flour in rice flour, allergens in flour, flour raw materials, such as eggs, wheat, milk, It can be suitably used as a simple discrimination technique or simple detection technique for dairy products, buckwheat, peanuts, and the like.
Moreover, the quality control method and flour product of the flour of the present invention can be managed to be a grain product having a predetermined quality, and as a result, a flour product having a predetermined quality can be provided. Moreover, the quality control method and product of flour of the present invention can be managed to be a grain product having a predetermined quality that does not contain an allergen, and as a result, provide a flour product having a predetermined quality that does not contain an allergen. Can do.

It is an example of the flowchart of the discrimination method of the flour of this invention. It is an example of the flowchart of the fluorescence intensity measurement process of the EEM information acquisition process of the discrimination method of flour shown in FIG. It is a block diagram of one embodiment of a flour discrimination device of the present invention. It is a block diagram of one Embodiment of the spectral illumination part of a measurement unit of the flour discrimination | determination apparatus shown in FIG. 3, and a fluorescence measurement part. It is an example of the fluorescence intensity measurement flowchart in the measurement unit shown in FIG. (A) And (b) is a figure showing EEM of the flour for pasta and the flour for bread of Example 1 of the present invention, respectively. It is a graph of the fluorescence characteristic which shows the result of the principal component analysis of Example 1 of this invention. It is a graph of the fluorescence characteristic according to the known flour product used in Example 1 of this invention. (A) And (b) is a figure showing EEM of the coarse powder and fine powder of the dry noodle of Example 2 of this invention, respectively. It is a graph of the fluorescence characteristic which shows the result of the principal component analysis of Example 2 of this invention. It is a graph of the fluorescence characteristic according to cereal flour of the dry noodles of the known particle size used in Example 2 of the present invention. (A) And (b) is a figure showing EEM of 100% oat flour of Example 3 of the present invention, and 30% wheat flour combination oat flour, respectively. It is a graph of the fluorescence characteristic which shows the result of the principal component analysis of Example 3 of this invention. It is a graph of the fluorescence characteristic according to the known different flour used in Example 3 of the present invention. It is a graph of the fluorescence characteristic which shows the result of the principal component analysis of Example 4 of this invention. It is a graph which shows the result of the cluster analysis of Example 5 of this invention. It is a graph which shows the result of discriminant analysis by the known sample of Example 6 of this invention. It is a table | surface which shows the result of the discriminant analysis of Example 6 of this invention. It is a graph which shows the result of the regression analysis of Example 7 of this invention. It is a schematic diagram explaining the neural network of this invention. It is a table | surface which shows the result by the neural network of Example 8 of this invention. It is a graph which shows the result of the principal component analysis of Example 9 of this invention.

10 Flour discrimination device (discrimination device)
DESCRIPTION OF SYMBOLS 12 Measurement unit 14 Flour discrimination | determination unit 16 Operation part 18 Display 20 Grain size acquisition part 21 Measuring object flour support part (support part)
22 Spectral Illumination Unit 24 Fluorescence Measurement Unit 26 Information Acquisition Unit 28 Fluorescence Characteristic Extraction Unit 30 Fluorescence Characteristic Verification Unit 32 Memory (DB)
34 Control unit 36 External memory (DB)
38 Light source 40, 48 Spectroscopic device 42 Excitation wavelength adjustment unit 44 Optical device 46 Excitation light irradiation unit 50 Fluorescence measurement device MF Measurement object (measuring flour)

Claims (20)

According to a predetermined characteristic of the flour obtained by pulverizing the grain, a flour discrimination method for discriminating the type of flour to be measured,
A particle size acquisition step of acquiring a particle size of the flour;
Measure the fluorescence intensity of the fluorescence generated from the flour irradiated with excitation light while changing the excitation wavelength irradiated to the flour and the fluorescence wavelength to be measured stepwise in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range. Excitation / fluorescence matrix information acquisition step of acquiring excitation / fluorescence matrix information of the flour,
Fluorescence characteristic extraction step of performing statistical analysis processing on the excitation / fluorescence matrix information of the flour obtained in this information obtaining step, and parameterizing and extracting the fluorescence property of the flour;
For the known types of flour, the particle size acquisition step, the excitation / fluorescence matrix information acquisition step and the fluorescence characteristic extraction step are performed in advance to acquire the particle size and the parameterized fluorescence characteristics, and the flour of the flour A reference fluorescence characteristic acquisition / storage step in which a known type and the obtained granularity are associated with the parameterized fluorescence characteristics and stored in a database in advance,
Search the database using the particle size of the flour to be measured acquired in the particle size acquisition step and the parameterized fluorescence characteristics of the flour to be measured extracted in the fluorescence property extraction step And depending on the particle size of the flour to be measured, the parameterized fluorescence characteristics of the flour to be measured and the flour for the known type of flour previously stored in the database And a fluorescence characteristic matching step of comparing the parameterized fluorescence characteristics to determine the type of the flour to be measured.   The method for discriminating flour according to claim 1, wherein the fluorescence characteristic extraction step acquires the fluorescence characteristic of the flour as a two-dimensional parameter from the excitation / fluorescence matrix information.   The said fluorescent characteristic extraction process performs the said statistical analysis process by multivariate analysis or data mining, The discrimination method of the flour of Claim 1 or 2 characterized by the above-mentioned.   The multivariate analysis or the data mining is performed using at least one method selected from the group consisting of data structure analysis, discriminant analysis, pattern classification, multi-factor data analysis, regression analysis, and a learning machine. 3. A method for discriminating flour according to 3.   5. The flour discrimination method according to claim 4, wherein the data structure analysis is performed using at least one method selected from the group consisting of principal component analysis, factor analysis, correspondence analysis, and independent component analysis.   In the fluorescence characteristic extraction step, the multivariate analysis or the data mining is performed using the principal component analysis method, and two or less principal component scores are extracted as parameters from the excitation / fluorescence matrix information, Is obtained as a fluorescence characteristic of the flour, The flour discrimination method according to claim 5.   The flour discrimination method according to claim 4, wherein the discriminant analysis is performed using a linear discriminant analysis method or a nonlinear discriminant analysis method.   In the fluorescence characteristic extraction step, the multivariate analysis or the data mining is performed using the discriminant analysis method, and the excitation / fluorescence matrix information is applied to a discriminant stored in the database in advance. The method for determining flour according to claim 7, further comprising: comparing the parameters stored in the database in advance and determining the type of the flour.   The method according to claim 4, wherein the pattern classification is performed by cluster analysis or multidimensional scaling.   The method for discriminating flour according to claim 4, wherein the regression analysis is performed using a linear regression, nonlinear regression, or multiple regression analysis technique.   In the fluorescence characteristic extraction step, the multivariate analysis or the data mining is performed using the method of the multiple regression analysis, and the excitation / fluorescence matrix information is applied to a calibration curve stored in advance in the database. The method according to claim 10, wherein the parameter is calculated and collated with a parameter stored in advance in the database.   The method according to claim 4, wherein the learning machine performs at least one method selected from the group consisting of a neural network, a self-organizing map, group learning, and a genetic algorithm. The flour is flour containing a specific component of the grain as a main component,
The fluorescence characteristic collation step is obtained in advance according to the fluorescence characteristics of the flour extracted in the fluorescence property extraction step, the flour for each component of the flour and the particle size thereof, and stored in the database. Collating with the parameterized fluorescence characteristics of flour containing each component of grain as a main component, and specifying that the flour to be measured is flour containing a specific component of the grain as a main component The method for discriminating flour according to any one of claims 1 to 12.   14. The grain flour according to claim 1, wherein the flour includes at least one selected from the group consisting of wheat flour, barley flour, rye flour, rice flour, corn flour, oat flour and beans flour and mixtures thereof. The discrimination method of the flour in any one.   15. The method according to claim 14, wherein the flour includes at least one selected from the group consisting of durum flour, strong flour, medium flour, and thin flour. The flour to be measured is flour containing allergens therein,
The method for determining flour according to any one of claims 1 to 15, wherein the type of the flour to be measured and the allergen contained in the flour are specified. The flour to be measured is a mixture of various types of flour,
The classification method of the flour in any one of Claims 1-16 by which the classification of the said flour used as the said measuring object and the specific flour contained in this flour are specified.   The said particle size acquisition process acquires the known particle size measured beforehand of the said flour used as the said measuring object as information, The discrimination method of the flour in any one of Claims 1-17. According to a predetermined characteristic of the flour obtained by pulverizing the grain, a flour discrimination device for discriminating the type of flour to be measured,
A particle size obtaining means for obtaining a particle size of the flour;
Spectral illumination means for irradiating the flour with the excitation light while changing the excitation wavelength stepwise in a predetermined excitation wavelength range;
According to the stepwise change in the excitation wavelength of the excitation light irradiated onto the flour by the spectral illumination means, the excitation light is changed in a predetermined fluorescence wavelength range while changing the fluorescence wavelength to be measured stepwise. Information acquisition means for measuring fluorescence intensity of fluorescence generated from the irradiated flour and obtaining excitation / fluorescence matrix information of the flour;
Fluorescence characteristic extraction means for performing a statistical analysis process on the excitation and fluorescence matrix information of the flour acquired by the information acquisition means, and parameterizing and extracting the fluorescence characteristics of the flour;
The parameterized fluorescence characteristics of the flour whose type and particle size are previously acquired using the spectral illumination unit, the information acquisition unit and the fluorescence characteristic extraction unit, and the known type and particle size of the flour. And a database for storing in advance in association with
Using the known particle size of the flour to be measured and the parameterized fluorescence characteristics of the flour to be measured extracted by the fluorescence characteristic extraction means, the database is searched and becomes the measurement object The parameterization of the flour for the known type of the flour having a particle size corresponding to the parameterized fluorescence characteristic of the flour and the particle size corresponding to the known particle size of the flour to be measured, which is stored in advance in the database A flour discrimination device, comprising: a fluorescence property matching means for checking the type of the flour to be measured by checking the fluorescence properties.   Grain product quality characterized by discriminating that it is a flour product having a predetermined quality using the flour discrimination method according to any one of claims 1 to 18 or the discrimination device according to claim 19. Management method.