JP2012098244A - Component distribution analysis method, component distribution analysis apparatus and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a component distribution analysis method, component distribution analysis apparatus and program in which, in component distribution analysis for visualizing the distribution of a target component from an obtained fluorescent fingerprint, the target component can be further identified and the component can be quantitatively evaluated.SOLUTION: While changing an excitation wavelength to be radiated and a fluorescent wavelength to be observed step by step within a predetermined excitation wavelength range and a predetermined fluorescent wavelength range, a fluorescent intensity is measured at a plurality of measuring points on a measurement target, fluorescent fingerprint information at each measuring point on the measurement target is acquired, and the fluorescent fingerprint information of the measurement target acquired for each measuring point is comparted with fluorescent fingerprint information of a target component acquired beforehand to calculate a similarity index. On the basis of the similarity index, a color to color each measuring point is determined and color mapping of the color to color each measuring point is performed, and an image is acquired in which the component included in the measurement target is visualized.

Description

  The present invention relates to a component distribution analysis method, a component distribution analysis device, and a program.

  Since the distribution state of the component is greatly related to the property of the measurement object as a sample, many techniques for visualizing the target component distribution have been developed. For example, in general, a method of staining and observing with an optical microscope or a method of observing with an electron microscope after freezing and drying is used to visualize the distribution of a specific component in a measurement object.

  However, in the conventional technique for visualizing the target component distribution, it is necessary to perform pretreatment such as staining, freezing, and drying of the measurement object. Since these pretreatment operations affect the measurement object, sampling for measurement is required. In addition, the pretreatment requires a lot of time and skill, and further, the measurement object. There was a possibility of altering things. In addition, since it is very difficult to uniformly stain the measurement object in the pretreatment, there is a problem that it is very difficult to obtain quantitativeness.

  Therefore, in recent years, there has been a demand for the development of a technique for visualizing the target component distribution more easily and non-destructively without performing pretreatment.

  For example, in the component distribution visualization method described in Patent Document 1 by the present applicant, the excitation wavelength irradiated on the measurement object and the fluorescence wavelength observed from the measurement object are not stained without staining the measurement object. Excitation / excitation matrix (EEM) (also known as “fluorescence fingerprint”) acquired by measuring fluorescence intensity while changing in stages is compressed by statistical analysis to extract characteristic fluorescence characteristics Then, by performing color mapping using colors corresponding to the extracted fluorescence characteristics, a plurality of components of the measurement object are simultaneously displayed as images.

  Here, the “fluorescence fingerprint” described in Patent Document 1 and Non-Patent Document 1 is also referred to as an excitation-fluorescence matrix (EEM), and the wavelength of the excitation light that irradiates the measurement object continuously. It means three-dimensional data obtained by measuring a fluorescence spectrum while changing it periodically. Since the shape of the fluorescent fingerprint is determined in a component-specific manner like a fingerprint, the measurer can detect subtle differences in components that cannot be determined by fluorescent measurement at a single single wavelength based on the fluorescent fingerprint. it can. In addition to the fluorescent fingerprint information, the measurer also includes the fluorescent fingerprint for measuring the fluorescent fingerprint for each pixel or each pixel block with positional information (that is, spatial information indicating the position of each pixel in the image). By using imaging, it is possible to visualize the distribution of specific components in the measurement object.

Japanese Patent No. 3706914

Susumu Shimoyama, Yuko Noda, “Non-destructive measurement of dyes used in ancient dyeing artifacts in three-dimensional fluorescence spectra”, Analytical Chemistry, 1992

  However, in the conventional component distribution visualization method (for example, Patent Document 1 and Non-Patent Document 1), the target component distribution can be visualized more easily and non-destructively without performing preprocessing. In order to confirm whether or not the component distribution obtained is the target component, there is a problem that the identification of the component is not considered. In addition, there is a problem that it is not guaranteed until quantitative evaluation of the component such as how much the target component is.

  The present invention has been made in view of the above problems, and in component distribution analysis for visualizing the distribution of the target component from the obtained fluorescent fingerprint, the target component can be further identified, and the component An object of the present invention is to provide a component distribution analysis method, a component distribution analysis device, and a program that can quantitatively evaluate.

  In order to achieve such an object, the component distribution analysis method of the present invention is a component distribution analysis method for analyzing an object to be measured, and includes an excitation wavelength to be irradiated in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range, and Fluorescence fingerprint information acquisition for measuring fluorescence intensity at a plurality of measurement locations on the measurement object and acquiring fluorescence fingerprint information at each measurement location on the measurement object while gradually changing the fluorescence wavelength to be observed Step, comparing the fluorescence fingerprint information of the measurement object acquired for each measurement location in the fluorescence fingerprint information acquisition step with the fluorescence fingerprint information of the target component acquired in advance, and calculating a similarity index A similarity index calculation step to calculate, a color determination step to determine a color to be colored in each measurement location based on the similarity index calculated in the similarity index calculation step, and the measurement location to the measurement location The color to be painted determined at decision step and color mapping, component distribution visualizing step of acquiring an image visualizing the components included in the object to be measured, characterized in that it comprises a.

  Further, the component distribution analysis method of the present invention is the component distribution analysis method described above, wherein the image acquired in the component distribution visualization step and another component that visualizes a component other than the component of the image are visualized. And a composite image creation step of creating a composite image in which a plurality of components included in the measurement object are simultaneously visualized by combining the images.

  The component distribution analysis method of the present invention is the component distribution analysis method described above, wherein the similarity index includes cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, and Morisita similarity. At least one of exponent, standard Euclidean distance similarity, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity, Jackard coefficient similarity, Dice coefficient similarity, and Simpson coefficient similarity It is characterized by being.

  The component distribution analysis method of the present invention is characterized in that in the component distribution analysis method described above, the component is a component derived from a living body.

  Moreover, the component distribution analysis method of the present invention is the component distribution analysis method described above, wherein the biological component includes at least one selected from the group consisting of protein, starch, polysaccharides, minerals, and lipids. It is characterized by that.

  In addition, the component distribution analyzer of the present invention includes the spectroscopic illumination device that irradiates the measurement target with a predetermined excitation wavelength and the spectroscopic imaging device that observes the measurement target with a predetermined fluorescence wavelength. A component distribution analyzer for analyzing a component distribution of an object, wherein the irradiation excitation wavelength and the observation fluorescence wavelength are changed stepwise in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range on the measurement object. Fluorescence fingerprint information acquisition unit for measuring fluorescence intensity at a plurality of measurement locations and acquiring fluorescence fingerprint information of each measurement location on the measurement object, acquired for each measurement location by the fluorescence fingerprint information acquisition unit A similarity index calculator that calculates the similarity index by comparing the fluorescent fingerprint information of the measurement object with the fluorescence fingerprint information of the target component acquired in advance, and calculated by the similarity index calculator A color determining unit that determines a color to be colored at each measurement location based on the similarity index; and color mapping the color to be colored determined by the color determination unit at each measurement location; A component distribution visualization unit that acquires an image in which the component included in the object is visualized is provided.

  The component distribution analyzer of the present invention is the component distribution analyzer described above, wherein the image acquired by the component distribution visualization unit and another image obtained by visualizing a component other than the component of the image. And a synthesized image creating unit that creates a synthesized image in which a plurality of components included in the measurement object are visualized simultaneously.

  In the component distribution analyzer of the present invention, the similarity index includes cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, and Morisita similarity. At least one of exponent, standard Euclidean distance similarity, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity, Jackard coefficient similarity, Dice coefficient similarity, and Simpson coefficient similarity It is characterized by being.

  Moreover, the component distribution analyzer of the present invention is characterized in that in the component distribution analyzer described above, the component is a component derived from a living body.

  The component distribution analyzer of the present invention is the component distribution analyzer described above, wherein the biological component includes at least one selected from the group consisting of protein, starch, polysaccharides, minerals, and lipids. It is characterized by that.

  The program of the present invention includes a spectral illumination device that irradiates a measurement object with a predetermined excitation wavelength, and a spectroscopic imaging device that observes the measurement object with a predetermined fluorescence wavelength. A program for causing a component distribution analyzer to analyze a distribution, wherein in the component distribution analyzer, the excitation wavelength to be irradiated and the fluorescence wavelength to be observed are stepwise in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range. A fluorescence fingerprint information acquisition step of measuring fluorescence intensity at a plurality of measurement locations on the measurement object and acquiring fluorescence fingerprint information at each measurement location on the measurement object, The similarity index is calculated by comparing the fluorescent fingerprint information of the measurement object acquired for each measurement location in the process with the fluorescent fingerprint information of the target component acquired in advance. A similarity index calculation step, a color determination step for determining a color to be colored at each measurement location based on the similarity index calculated in the similarity index calculation step, and the color determination at each measurement location A component distribution visualization step of performing color mapping of the color to be colored determined in the step and acquiring an image in which the component included in the measurement object is visualized is executed.

  Further, the program of the present invention synthesizes the image acquired in the component distribution visualization step and another image obtained by visualizing a component different from the component of the image in the program described above. And a composite image creating step of creating a composite image in which a plurality of components included in the measurement object are simultaneously visualized.

  In the program according to the present invention, the similarity index includes cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, Morisita similarity index, and standard Euclidean distance similarity. At least one of degree, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity, Jackard coefficient similarity, Dice coefficient similarity, and Simpson coefficient similarity To do.

  The program of the present invention is characterized in that, in the above program, the component is a component derived from a living body.

  The program of the present invention is characterized in that, in the above program, the biological component includes at least one selected from the group consisting of protein, starch, polysaccharide, mineral, and lipid.

  As described above, the present invention measures the fluorescence fingerprint for each pixel of the measurement object using the fluorescence fingerprint imaging that combines the fluorescence fingerprint and the imaging technique, and examines the similarity to the fluorescence fingerprint of the target component alone. The distribution of the target component, which is the target component, is quantitatively visualized and evaluated by converting the similarity to a color.

  According to the present invention, in a predetermined excitation wavelength range and a predetermined fluorescence wavelength range, while changing the excitation wavelength to be irradiated and the fluorescence wavelength to be observed in a stepwise manner, By measuring the fluorescence intensity using the pixel block unit as a measurement location, the fluorescence fingerprint information of each measurement location on the measurement target is acquired, and the fluorescence fingerprint information of the measurement target acquired for each measurement location is acquired in advance. The target component (target component) is compared with the fluorescence fingerprint information, and an index corresponding to the similarity between them (ie, the similarity index) is calculated, and each measurement is performed based on the calculated similarity index. By determining the color to be colored at the location and color mapping the color to be colored determined at each measurement location, an image in which the similarity index is visualized by color is acquired. As a result, the target component can be identified at each measurement location, and the component can be evaluated quantitatively. Thereby, this invention has an effect that the distribution of the target component specified beforehand can be visualized with quantitativeness. In addition, even if there are a plurality of target components, the present invention measures the fluorescence fingerprint imaging of a measurement object if the fluorescence fingerprints of a plurality of target components acquired in advance are known, and the similarity is obtained. Since the calculation is performed for each component, the distribution of a plurality of target components can be visualized separately and quantitatively.

  In addition, according to the present invention, a plurality of images included in the measurement object are obtained by synthesizing the acquired visualized image of a target component and another image obtained by visualizing a component different from the component of the image with a different color. Since a composite image in which the components are visualized at the same time is created, the distribution of the visualized components is evaluated in a superimposed state, and the relationship between the components can be examined.

  Further, according to the present invention, the similarity index includes cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, Morisita similarity index, standard Euclidean distance similarity, Mahalanobis distance similarity, At least one of Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity, Jackard coefficient similarity, Dice coefficient similarity, and Simpson coefficient similarity can be used. Thereby, the present invention can more accurately identify the target component by replacing the comparison of the fluorescence fingerprints composed of multivariate parameters with one continuous index called the similarity index, and the component There is an effect that the quantitative property can be more accurately obtained.

  Further, according to the present invention, in principle, any component that can acquire a fluorescent fingerprint can be visualized. The component having a fluorescent fingerprint may be a protein, starch, polysaccharide, mineral, lipid, or a composite thereof, and other biological components, and includes at least one selected from the group consisting of these components. There is an effect that distribution analysis can be performed.

FIG. 1 is a diagram showing an example of a fluorescent fingerprint as a contour graph of three-dimensional data. FIG. 2 is a bird's-eye view showing an example of the fluorescent fingerprint of FIG. FIG. 3 is a flowchart showing the processing of the component distribution analysis method of the present invention. FIG. 4 is a diagram illustrating an example of a fluorescence image. FIG. 5 is a diagram illustrating an example of a fluorescent fingerprint in each pixel of a fluorescent image of a measurement object. FIG. 6 is a diagram illustrating an example of a fluorescent fingerprint of a target component acquired in advance. FIG. 7 is a diagram illustrating an example of cosine similarity. FIG. 8 is a diagram illustrating an example of cosine similarity. FIG. 9 is a diagram illustrating an example of cosine similarity. FIG. 10 is a diagram illustrating an example of fluorescent fingerprints and color axes of different similarity indices. FIG. 11 is a diagram illustrating an example of an image obtained by obtaining a similarity index of a target component gluten with a fluorescent fingerprint for a measurement target gluten and visualizing the gluten. FIG. 12 is a diagram illustrating an example of an image obtained by obtaining a similarity index with a fluorescent fingerprint of a target component gluten for a measurement object bread dough and visualizing gluten in the bread dough. FIG. 13 is a diagram illustrating an example of an image obtained by obtaining a similarity index with a fluorescent fingerprint of a target component gluten for a measurement target starch and visualizing gluten in starch. FIG. 14 is a diagram showing another example of an image obtained by obtaining a similarity index with a fluorescent fingerprint of the target component starch for the measurement target gluten and visualizing starch in gluten. FIG. 15 is a diagram illustrating another example of an image obtained by obtaining a similarity index with a fluorescent fingerprint of a target ingredient starch for a measurement object bread dough and visualizing starch in the bread dough. FIG. 16 is a diagram illustrating another example of an image in which starch is visualized by obtaining a similarity index with the fluorescent fingerprint of the target component starch for the measurement target starch. FIG. 17 is a diagram showing an example of a composite image obtained by synthesizing the image of bread dough visualized with gluten in FIG. 12 and the image of bread dough visualized with starch in FIG. FIG. 18 is a block diagram showing an example of the configuration of the component distribution analysis system of the present embodiment. FIG. 19 is a block diagram illustrating an example of a fluorescent image photographing device. FIG. 20 is a perspective view showing an example of a fluorescent image photographing device. FIG. 21 is a diagram showing an example of a stained image of gluten in which gluten is stained red with rhodamine B. FIG. 22 is a diagram illustrating an example of a stained image of starch obtained by staining starch green with FITC. FIG. 23 is a diagram showing an example of a synthetic stained image obtained by synthesizing the gluten stained image of FIG. 21 and the starch stained image of FIG. FIG. 24 is a diagram showing a complementary relationship between the cosine similarity with starch and the cosine similarity with gluten for the composite image in FIG. 17. FIG. 25 is a diagram showing a complementary relationship between the fluorescence intensity of starch and the fluorescence intensity of gluten in the synthetic stained image of FIG. FIG. 26 is a diagram illustrating an example of a composite image of bread dough at the stage of insufficient mixing. FIG. 27 is a diagram illustrating an example of a composite image of bread dough at the optimum mixing stage. FIG. 28 is a diagram showing an example of a composite image of bread dough at an excessive mixing stage. FIG. 29 is a diagram illustrating an example of component distribution quantification processing when the component distribution is non-uniform. FIG. 30 is a diagram illustrating another example of the component distribution quantification process when the component distribution is non-uniform. FIG. 31 is a diagram illustrating an example of a component distribution quantification process when the component distribution is uniform. FIG. 32 is a diagram illustrating another example of the component distribution quantification process when the component distribution is uniform. FIG. 33 is a diagram illustrating an example of the result of quantifying the component distribution. FIG. 34 is a diagram illustrating an example of uniformity of component distribution in each mixing stage. FIG. 35 is a diagram showing an example of the bubble area ratio in each mixing stage.

  Examples of preferred embodiments of [I] component distribution analysis method, [II] component distribution analysis system, and program according to embodiments of the present invention will be described below in detail with reference to FIGS. . Note that the present invention is not limited to the embodiments.

[I] Component Distribution Analysis Method First, the fluorescence fingerprint will be described with reference to FIGS. 1 and 2. Here, FIG. 1 is a diagram showing an example of a fluorescent fingerprint as a contour graph of three-dimensional data. FIG. 2 is a bird's-eye view showing an example of the fluorescent fingerprint of FIG.

In the present embodiment, as shown in FIG. 1, the “fluorescence fingerprint” refers to an excitation wavelength λ _Ex irradiating the measurement object, a fluorescence wavelength λ _{Em of} light emitted from the measurement object, and a fluorescence intensity I _Ex of the measurement object. _{, Em} is a contour graph of three-dimensional data composed of three axes.

  Further, as shown in FIG. 2, the “fluorescence fingerprint” can be represented in a plane by plotting the fluorescence intensity at each point in a contour line with the horizontal axis representing the fluorescence wavelength and the vertical axis representing the excitation wavelength. Fluorescent fingerprints have the advantages of being able to characterize the measurement object without pretreatment such as fluorescent staining, being easy to operate and measuring in a short time, and being more sensitive than UV absorbance. Therefore, it is a widely used technique for analyzing the internal structure of foods, measuring suspended matters in the atmosphere, and specifying dye raw materials.

  In this embodiment, this component distribution analysis method uses this fluorescence fingerprint to measure and extract only the fluorescence information specific to the target component from a three-dimensional enormous amount of information (data amount). In this method, the component distribution inside the substance is visualized, and further, the component is identified (identified) to quantitatively evaluate the component (for example, to obtain the quantitative property of the component).

  Hereinafter, as an example, the process of visualizing the distribution of gluten and / or starch in the bread dough by measuring and analyzing the bread dough as a measurement object based on the fluorescence fingerprint, as appropriate according to the flowchart of FIG. This will be described with reference to FIG.

  As shown in FIG. 3, in this component distribution analysis method, first, a measurement object for analyzing a component distribution is prepared by a measurer as a premise (step S1).

  Here, in the present embodiment, the “measurement object” means a sample that is a measurement object when analyzing a component distribution using a fluorescent fingerprint. As an example, the measurement object may include at least a flour product (for example, bread dough, noodles such as udon and soba, and confectionery such as cookies and cakes). In the present embodiment, the measurement object is not limited to these, and may be another sample that can analyze the component distribution using a fluorescent fingerprint.

  In the present embodiment, “component” means a component contained in the measurement object. As an example, the component may be a component derived from a living body, for example. In addition, the biological component may include, for example, at least one selected from the group consisting of proteins (eg, gluten), starch (starch), polysaccharides, minerals, and lipids. In addition, in this embodiment, a component is not limited to these, The composite of these components and / or other components derived from a living body may be sufficient.

  Then, in this component distribution analysis method, in the fluorescent fingerprint information acquisition step, the fluorescent fingerprint of the measurement object prepared in step S1 is measured for each pixel (step S2). That is, in the fluorescence fingerprint information acquisition step of step S2, a plurality of measurements on the measurement object are performed 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. The fluorescence intensity at each location (that is, each pixel unit or pixel block unit constituting the fluorescence image of the measurement object) is measured, and fluorescence fingerprint information of each measurement location on the measurement object is acquired.

  Here, with reference to FIG. 4, an example of the fluorescence fingerprint information acquisition process of step S2 is demonstrated. Here, FIG. 4 is a diagram illustrating an example of the fluorescence image.

  For example, as shown in FIG. 4, in the fluorescent fingerprint information acquisition step of step S2, the fluorescent fingerprint is measured for each 1344 × 1024 pixels (pixels) within the measurement range of 63 fluorescent images taken, for example. To do. At this time, the measurer changes the combination of the excitation wavelength (for example, 7 wavelengths at intervals of 10 nm within the range of 260 to 320 nm) and the fluorescence wavelength (for example, 9 wavelengths at intervals of 10 nm in the range of 370 to 450 nm) A fluorescent image of the measurement object (for example, bread dough) is taken under a total of 63 wavelength conditions. Here, if 63 luminance values are extracted by skewing in the direction of the captured image for one pixel and plotted on the excitation wavelength axis and the fluorescence wavelength axis, the fluorescence fingerprint at that pixel can be drawn as shown in FIG. it can.

  That is, in the fluorescence fingerprint information acquisition step of step S2, a fluorescence image is taken with respect to the measurement object under 63 wavelength conditions in which the combination of the excitation wavelength to be irradiated and the fluorescence wavelength to be observed is different, and the fluorescence intensity of each pixel in each fluorescence image From the above, the fluorescence fingerprint information at each measurement location on the measurement object is acquired.

  Returning to FIG. 3, in this component distribution analysis method, in the similarity index calculation step, the fluorescent fingerprint of the measurement object (for example, bread dough) acquired for each measurement location in the fluorescent fingerprint information acquisition step in step S <b> 2. The information is compared with the fluorescence fingerprint information of the target component (for example, gluten, starch, etc.) acquired in advance, and a similarity index is calculated (step S3).

  Here, in the present embodiment, the “similarity index” refers to a fluorescent fingerprint of a sample to be measured (for example, bread dough as a measurement target) and a fluorescent fingerprint of a target component (for example, gluten, starch, etc.). This means a scale indicating the similarity of fingerprint patterns. For example, the similarity index includes cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, Morisita similarity index, standard Euclidean distance similarity, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev It may be at least one of a distance similarity, a Minkowski distance similarity, a Jackard coefficient similarity, a Dice coefficient similarity, and a Simpson coefficient similarity. In the present embodiment, the similarity index is not limited to these, and may be another index indicating the similarity of the fluorescent fingerprint pattern.

  Here, with reference to FIG. 5 and FIG. 6, an example in the case of calculating cosine similarity as a similarity index in the similarity index calculation step of step S3 will be described. Here, FIG. 5 is a diagram illustrating an example of the fluorescence fingerprint in each pixel of the fluorescence image of the measurement object. FIG. 6 is a diagram illustrating an example of a fluorescent fingerprint of a target component acquired in advance.

  For example, in the similarity index calculation process of step S3, in order to identify the component from each pixel of the fluorescence fingerprint, the fluorescence of the measurement object whose component acquired in the fluorescence fingerprint information acquisition process of step S2 is unknown is known. Analysis is performed by paying attention to the similarity between the fingerprint (see FIG. 5) and the fluorescence fingerprint (see FIG. 6) of the target component (target component) acquired in advance. Here, as an example, the measurement object is one pixel of bread dough, and the target components are gluten and starch. That is, in the similarity index calculation process of step S3, the amount or existence probability of the target component (for example, gluten, starch, etc.) contained in the measurement target (for example, bread dough) and the measurement target and the target component Under the premise that there is a correlation between the similarity of fluorescent fingerprints, the similarity of fluorescent fingerprints (for example, cosine similarity) is calculated.

  Here, in this embodiment, “cosine similarity” is similarity based on the cosine of an angle formed by two vectors (for example, vector x and vector y), as shown below, and the inner product of the vectors is The value divided by the product of the lengths of the two vectors.

Cosine similarity = Cosθ of vector x and vector y
= Vector inner product / Vector length = x · y / (| x | * | y |)
= (X ₁ y ₁ + x ₂ y ₂ + ... + x _n y _n ) / ((x ₁ ² + x ₂ ² + ... + x _n ² ) * (y ₁ ² + y ₂ ² + ... + y _n ² )) ^1/2

Here, an example of the cosine similarity will be described with reference to FIGS. As shown in “Cos θ ₁ = −1” in FIG. 7, the minimum value of −1 indicates that the directions of two vectors (for example, vector x and vector y) are reversed and the similarity is low. Conversely, as shown in “Cos θ ₂ = 1/2” in FIG. 8 and “Cos θ ₃ = 1” in FIG. 9, the closer the maximum value is to 1, the more the directions of the two vectors (vector x, vector y) are the same. We show that the similarity is high. 7 to 9 exemplify the case of a two-dimensional vector in the case of n = 2 for the sake of explanation, in the present embodiment, since there are 63 wavelength conditions, n = 63 is assumed. The similarity of 63-dimensional fluorescent fingerprint information measured by the same calculation method can be calculated.

  Returning to FIG. 3, in this component distribution analysis method, in the color determination step, the color of each pixel is determined from the similarity (eg, cosine similarity) in step S3 (step S4). That is, in the color determination process in step S4, each measurement location should be colored based on the similarity index calculated in the similarity index calculation process in step S3 (for example, the cosine similarity shown in FIGS. 7 to 9). Determine the color.

  Here, with reference to FIG. 10, an example of the color determination process in step S4 when the cosine similarity is calculated as the similarity index in the similarity index calculation process in step S3 will be described. Here, FIG. 10 is a diagram illustrating an example of fluorescent fingerprints and color axes of different similarity indices.

For example, as shown in FIG. 10, in the color determination process in step S4, the cosine similarity calculated in the similarity index calculation process in step S3 (for example, Cos θ ₁ to Cos θ ₃ corresponding to FIGS. 7 to 9). In order to visually express the value of cosine, by using a color axis as shown in FIG. 10, the magnitude of cosine similarity (Cos θ ₁ <Cos θ ₂ <Cos θ ₃ in FIG. 10) and the brightness of the color axis are expressed. Correspondingly, conversion to the color scale is performed. In this embodiment, as an example, the cosine similarity with gluten is associated with a red light and dark color, and the cosine similarity with starch is associated with a green light and dark color.

  Returning to FIG. 3, in this component distribution analysis method, an image is created by performing coloring according to the color determined in the color determination step in step S4 in the component distribution visualization step (step S5). That is, in the component distribution visualization process of step S5, the color to be colored determined in the color determination process of step S4 is color-mapped at each measurement location, and the components (for example, gluten, starch, etc.) contained in the measurement object Get an image that visualizes. Here, in the present embodiment, the image acquired in the component distribution visualization step may be a color image or a monochrome image.

  Here, an example of an image colored according to the color determined in the color determination step in step S4 in the component distribution visualization step in step S5 will be described with reference to FIGS.

  Here, FIG. 11 is a diagram illustrating an example of an image in which a gluten is visualized by obtaining a similarity index between the measurement target gluten and a fluorescent fingerprint of the target component gluten. FIG. 12 is a diagram showing an example of an image obtained by obtaining a similarity index with the fluorescence fingerprint of the target component gluten for the measurement object bread dough and visualizing the gluten in the bread dough. FIG. 13 is a diagram illustrating an example of an image obtained by obtaining a similarity index with a fluorescent fingerprint of the target component gluten for the measurement target starch and visualizing gluten in the starch. Moreover, FIG. 14 is a figure which shows another example of the image which calculated | required the similarity index with the fluorescence fingerprint of the target component starch with respect to the measurement object gluten, and visualized the starch in gluten. Moreover, FIG. 15 is a figure which shows another example of the image which calculated | required the similarity index with the fluorescence fingerprint of the target component starch with respect to the measurement object bread dough, and visualized the starch in bread dough. Moreover, FIG. 16 is a figure which shows another example of the image which calculated | required the similarity index with the fluorescence fingerprint of the target component starch with respect to the measuring object starch, and visualized starch.

  For example, in the component distribution visualization process in step S5, the measurement shown in FIGS. 11 to 16 is performed using the color applied to each pixel based on the color axis shown in FIG. 10 in the color determination process in step S4. An image in which components (for example, gluten, starch, etc.) contained in an object (for example, bread dough) are visualized is acquired.

  As an example, the images shown in FIGS. 11 to 13 are images drawn by associating colors with values of cosine similarity with the fluorescence fingerprint of gluten that is the target component. Here, the portion indicated by (i) in the gluten image in FIG. 11 indicates bubbles. In the image of bread dough in FIG. 12, the red portion indicated by (ii) indicates a gluten component. In addition, the starch image in FIG. 13 is an image drawn corresponding to the color of cosine similarity with the fluorescent fingerprint of gluten, which is the target component, so that the dark color indicating that the similarity between starch and gluten is low ( For example, black) (iii).

  As another example, the images shown in FIGS. 14 to 16 are images drawn by correlating colors with values of cosine similarity with the fluorescent fingerprint of starch, which is the target component. Here, since the gluten image in FIG. 14 is an image drawn corresponding to the color of cosine similarity with the fluorescent fingerprint of starch, which is the target component, the similarity between gluten and starch is low. It is represented by a dark color (for example, black) (iv). In the image of bread dough in FIG. 15, the green part indicated by (v) indicates a starch component. Moreover, the part which (vi) shows in the image of the starch of FIG. 16 shows a bubble.

  As described above, in this embodiment, in the color determination process in step S4 and the component distribution visualization process in step S5, the 63 wavelength condition (seven excitation wavelengths × 9 wavelengths) acquired in the fluorescent fingerprint information acquisition process in step S2 For each pixel of the fluorescence image of the measurement object (for example, bread dough) in the fluorescence wavelength condition) and the target object (for example, gluten, starch, etc.) calculated in the similarity index calculation step in step S3. Based on the similarity (for example, cosine similarity), a color corresponding to the similarity is applied to the pixel to create a pseudo image of the fluorescent fingerprint (see FIGS. 11 to 16).

  Here, in this embodiment, the color scale (color axis) used for the images shown in FIGS. 11 to 16 may be any combination of colors. For example, in the images shown in FIGS. 11 to 13, the gluten color is red, the portion with a low similarity index is black, and the portion with a high similarity index is red (see FIG. 12 (ii)). However, this is a combination of black and red to give similarity to a gluten-stained image (for example, an image in which gluten is stained red with rhodamine B (see FIG. 21 described later)). In the same way, in the images as shown in FIGS. 14 to 16, the starch color is green, and the portion having a low similarity index is black, A portion having a high similarity index is represented in green (see FIG. 15 (v)), and this is a stained image of starch (for example, an image in which starch is stained green by FITC (see FIG. 22 described later)). Kind of Are those using a combination of black and green in order to have sex, it may be a combination of any color in practice.

  Returning to FIG. 3, in this component distribution analysis method, next, in the composite image creation step, an image (for example, an image of bread dough visualized as gluten shown in FIG. 12) acquired in the component distribution visualization step in step S <b> 5. ) And another image (for example, an image of bread dough that visualizes starch shown in FIG. 15) in which a component (for example, starch) other than the component (for example, gluten) in the image is visualized, Then, a composite image in which a plurality of components (for example, gluten, starch, etc.) contained in the measurement object (for example, bread dough) are visualized simultaneously is created (step S6). Here, in the present embodiment, the composite image created in the composite image creation step may be a color image or a monochrome image.

  Here, an example of the composite image created in the composite image creation process in step S6 will be described with reference to FIG. Here, FIG. 17 is a diagram illustrating an example of a composite image obtained by synthesizing the bread dough image in which gluten is visualized in FIG. 12 and the bread dough image in which starch is visualized in FIG. In addition, in this embodiment, although the process of a synthetic image creation process is demonstrated about the image of bread dough which visualized gluten (refer FIG. 12) and the image of bread dough which visualized starch (refer FIG. 15), The processing is not limited to two types of components (for example, gluten, starch, etc.), and two or more images visualized with respect to a plurality of two or more types of components can be similarly synthesized.

  For example, in the composite image creation process in step S6, an image obtained from the cosine similarity between the gluten obtained in the component distribution visible process in step S5 and the fluorescent fingerprint of starch is synthesized, as shown in FIG. A composite image in which a plurality of components (for example, gluten, starch, etc.) contained in a measurement object (for example, bread dough) is simultaneously visualized is acquired. Here, in the composite image of FIG. 17, the distribution of gluten in the red part indicated by (i) and the distribution of starch in the green part indicated by (ii) are complementary. That is, in the composite image of FIG. 17, there is little starch in the portion with a lot of gluten (that is, the red portion shown in (i)), and there is little gluten in the portion with a lot of starch (ie, the green portion shown in (ii)). As a result, the actual distribution state is appropriately reflected. Here, the part indicated by (iii) in FIG. 17 indicates bubbles.

  In the present embodiment, the component distribution analysis method has been described by taking the distribution of the component of bread dough as an example of the measurement object. However, the measurement object, the measurement condition of the fluorescence fingerprint, the calculation method of the similarity index, the color mapping are described. Various conditions such as these can be selected as appropriate without departing from the object of the present invention.

[II] Component Distribution Analysis System Next, the configuration of the component distribution analysis system of the present invention will be described with reference to FIGS. Note that the fluorescence image capturing device and the component distribution analysis device provided in the component distribution analysis system (fluorescence fingerprint imaging system) in the embodiment of the present invention can be suitably used for the above-described component distribution analysis method. The measuring apparatus used for the component distribution analysis method is not limited to this.

  Here, FIG. 18 is a block diagram showing an example of the configuration of the component distribution analysis system of the present embodiment, and conceptually shows only the portion related to the present invention in the configuration.

  As shown in FIG. 18, the component distribution analysis system includes a fluorescence image capturing device 10 and a component distribution analysis device 20. The fluorescent image photographing device 10 is a device that acquires fluorescent fingerprint information, and includes a spectral illumination device 11 and a spectral photographing device 12. The component distribution analysis device 20 is a device that analyzes and visualizes the component distribution of the measurement object 13 from the fluorescence fingerprint information acquired by the fluorescence imaging device 10, and includes the memory 21, the control unit 23, and the calculation processing unit 24. The measurer inputs measurement conditions and the like to the component distribution analyzer 20 by using the keyboard / mouse 22.

  That is, in the present invention, the spectral illumination device 11 irradiates the measurement target 13 with a predetermined excitation wavelength, and the spectral imaging device 12 observes the measurement target 13 with a predetermined fluorescence wavelength. In addition, the component distribution analyzer 20 includes the spectral illumination device 11 and the spectral imaging device 12, and analyzes the component distribution of the measurement target 13.

  Here, FIG. 19 is a block diagram showing an example of the fluorescent image photographing device 10, and conceptually shows only a part related to the present invention in the configuration.

  As shown in FIG. 19, the fluorescence image photographing device 10 includes a spectral illumination device 11 and a spectral photographing device 12. The spectroscopic illumination device 11 is a device that irradiates the measurement target 13 with excitation light having a predetermined wavelength to generate fluorescence from the components of the measurement target 13. The spectral illumination device 11 has excitation wavelength variable means for arbitrarily changing the excitation wavelength to be irradiated.

  The spectroscopic imaging device 12 is a device that captures a fluorescence image of the measurement target 13 at a predetermined fluorescence wavelength and transmits image information to the component distribution analyzer 20. The spectroscopic imaging device 12 captures a fluorescence image by selectively capturing a specific fluorescence wavelength among the fluorescence emitted from the measurement object 13. The spectroscopic imaging device 12 has a fluorescence wavelength variable means for arbitrarily changing the observed fluorescence wavelength.

  Here, a perspective view showing an example of the fluorescence image capturing apparatus 10 of FIG. 20 will be described.

  In the example of FIG. 20, the spectral illumination device 11 includes a light source 110, a spectral device 112, an excitation wavelength adjustment unit 114, an optical fiber 116, and an excitation light irradiation unit 118. The spectral imaging device 12 includes an optical unit 120, a spectral device 122, and an image imaging device 124.

  As the light source 110 of the spectral illumination device 11, a xenon lamp, a tungsten lamp, a wavelength tunable laser, or the like that emits white light can be used. The light emitted from the light source 110 has a wide wavelength band from the ultraviolet region to the visible region, for example, and is split into excitation light having a predetermined wavelength in the spectroscopic device 112. The measurer can set the excitation light to an arbitrary wavelength by adjusting the excitation wavelength adjusting unit 114. When a wavelength tunable laser is used as the light source 110, since the excitation light having a desired wavelength is directly obtained from the light source 100, the spectroscopic device 112 is unnecessary.

  As the spectroscopic device 112, for example, a band-pass filter (BPF), an AOTF (Acoustic Optical Tunable Filter), a liquid crystal tunable filter, an interference filter, a diffraction grating, or the like can be used. Moreover, you may use the LED light source, DLP light source, etc. which integrated the light source 110 and the spectroscopy apparatus 112. FIG.

  The light converted into the excitation light having a predetermined wavelength in the spectroscopic device 112 is sent to the excitation light irradiation unit 118 via the optical fiber 116 and is uniformly irradiated to the measurement object 13 from the excitation light irradiation unit 118. By irradiating excitation light having a predetermined excitation wavelength, the measurement object 13 emits fluorescence with a component-specific pattern. Here, as the excitation light irradiation unit 118, for example, a rod lens or the like can be used.

  The measurement object 13 is enlarged to a size suitable for image capturing by the optical unit 120. The optical unit 120 is, for example, an objective lens, a microscope, a zoom lens, a macro lens, a wide angle lens, or the like. If the measurement object 13 has a sufficient size for taking an image, the image can be taken without using the optical unit 120.

  By irradiating the excitation light, the fluorescence emitted from the measurement object 13 is split into light having a predetermined wavelength by the spectroscopic device 122. The spectroscopic device 122 has means for arbitrarily setting the fluorescence wavelength to be observed. As the spectroscopic device 122, for example, a band-pass filter (BPF), an AOTF (Acoustic Optical Tunable Filter), a liquid crystal tunable filter, a PGP (prism / grating / prism), an interference filter, a diffraction grating, or the like can be used.

  By passing through the spectroscopic device 122, only light having a predetermined wavelength among the fluorescence emitted from the measurement object 13 is captured as an image by the image capturing device 124. As the image capturing device 124, a CCD camera, a scanning image camera, or the like that can capture a monochrome image and / or a color image can be used.

From the brightness of each pixel of the image obtained by the image photographing device 124, the fluorescence intensity at each measurement location on the measurement object 13 is determined. That is, for each measurement location on the measurement object 13, the fluorescence intensity at a specific excitation wavelength λ _Ex (irradiation wavelength) and a specific fluorescence wavelength λ _Em (observation wavelength) is determined from the brightness of the pixel at the corresponding position of each measurement location. Information on I _{Ex and Em} is obtained. Therefore, by using the fluorescence imaging apparatus 10, it is possible to measure the fluorescence intensity at a specific excitation wavelength and fluorescence wavelength at a time at all points (pixels) on the measurement object 13.

  When obtaining n-dimensional fluorescence fingerprint information at each point on the measurement object 13, n fluorescence images are taken while changing the combination of the excitation wavelength and the fluorescence wavelength. The image information of the n fluorescent images photographed by the image photographing device 124 is transferred to the component distribution analyzing device 20.

  The excitation wavelength to be irradiated and the fluorescence wavelength to be observed can be arbitrarily changed manually or automatically by the measurer. Preferably, the fluorescence imaging apparatus 10 includes means for automatically changing the excitation wavelength and the fluorescence wavelength.

  Here, returning to FIG. 18, the component distribution analyzer 20 will be described.

  As shown in FIG. 18, the component distribution analyzer 20 acquires fluorescent fingerprint information for each pixel from the image information of n fluorescent images captured by the image capturing device 124 of the spectroscopic capturing device 12 (FIG. 18). , Processing by the fluorescence fingerprint information acquisition unit 24-1). Next, the similarity index is calculated by comparing the fluorescence fingerprint information of the measurement object 13 acquired for each measurement location with the fluorescence fingerprint information of the target component acquired in advance stored in the memory 21 (FIG. 18). , Processing by the similarity index calculation unit 24-2). Next, the color to be colored at each measurement location is determined based on the calculated similarity index (processing by the color determination unit 24-3 in FIG. 18). Next, the color to be colored determined at each measurement location is color-mapped, and an image in which components included in the measurement target 13 are visualized is acquired (processing by the component distribution visualization unit 24-4 in FIG. 18). Next, the acquired image and another image obtained by visualizing a component different from the component of the image are combined to create a composite image in which a plurality of components included in the measurement target 13 are simultaneously visualized (FIG. 18). The process by the composite image creation unit 24-5).

  The memory 21 is transferred from the image capturing device 124 to the component distribution analysis device 20, and the image information acquired by the fluorescence fingerprint information acquisition unit 24-1 of the calculation processing unit 24 and the similarity index calculation unit 24 of the calculation processing unit 24. -2 stores the fluorescence fingerprint information of the target component acquired in advance, which is referred to as a control (comparison control) when calculating the similarity index by -2.

  The control unit 23 irradiates the spectral illumination device 11 with the excitation wavelength and the spectral imaging device 12 so as to capture the fluorescence image of the measurement target 13 in the excitation wavelength range, fluorescence wavelength range, and wavelength pitch input by the measurer. Gives an instruction to adjust the observed fluorescence wavelength, and instructs the calculation processing unit 24 to perform the process.

  The fluorescence fingerprint information acquired by the fluorescence fingerprint information acquisition unit 24-1 of the calculation processing unit 24 at a plurality of measurement locations on the measurement target 13 (that is, n fluorescence images transferred from the fluorescence imaging apparatus 10). Is stored in the memory 21 of the component distribution analyzer 20. When the measurer instructs the calculation processing unit 24 to perform processing through the keyboard / mouse 22, first, the similarity index calculation unit 24-2 of the calculation processing unit 24 performs n fluorescence stored in the memory 21. From the image information of the image, n-dimensional fluorescent fingerprint information is extracted for each pixel. Next, the similarity index calculation unit 24-2 of the calculation processing unit 24 includes n-dimensional fluorescent fingerprint information obtained for each pixel, and fluorescent fingerprint information of components of the measurement target 13 acquired in advance in the memory 21. And the similarity index is calculated. The calculated similarity index is stored in the memory 21.

  When the measurer inputs an instruction to perform color mapping to the control unit 23 through the keyboard / mouse 22, the control unit 23 displays the similarity index stored in the memory 21 in the color determination unit 24-3 of the calculation processing unit 24. Reading and instructing each pixel to designate a color to be colored based on the color index designated by the measurer. That is, the color determination unit 24-3 determines a color to be colored at each measurement location based on the similarity index calculated by the similarity index calculation unit 24-2.

  When the color to be colored for each pixel is determined, the color determination unit 24-3 of the calculation processing unit 24 transmits information on the color to be colored to each pixel to the display 30. By color-mapping each pixel on the display 30, the component distribution of the measurement target 13 is visualized. The component distribution visualization is not necessarily performed on the display 30, and any method that can visualize the component distribution of the measurement target 13 such as printing a component distribution visualization image with a printer may be used.

  That is, the component distribution visualization unit 24-4 of the calculation processing unit 24 color-maps the color to be colored determined by the color determination unit 24-3 at each measurement location, and visualizes the component included in the measurement object 13. Get an image.

  Further, in the calculation processing unit 24, the measurer performs processing (for example, processing from the similarity index calculation unit 24-2 to the composite image creation unit 24-5) with respect to the calculation processing unit 24 through the keyboard / mouse 22. When instructed, the composite image creation unit 24-5 of the calculation processing unit 24 synthesizes the image acquired by the component distribution visualization unit 24-4 and another image that visualizes a component other than the component of the image. Then, a composite image in which a plurality of components included in the measurement target 13 are visualized at the same time is created.

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

[Example 1]
(1. Purpose of the first embodiment)
The distribution of gluten and starch, which changes depending on the mixing stage of bread dough, is an important factor that affects the texture of bread. However, in order to visualize the distribution, it was necessary to perform pretreatment such as dyeing. On the other hand, hyperspectral imaging, which simultaneously visualizes spatial information on the sample plane and optical information at each point, has been attracting attention as a technique that enables non-destructive visualization without requiring pretreatment of the sample. In particular, fluorescence fingerprint imaging using fluorescence fingerprints in which fluorescence spectra at a plurality of excitation wavelengths are arranged three-dimensionally as optical information is a visualization method that makes use of the superiority of fluorescence fingerprints that enable discrimination of specific components. It can be said. The purpose of Example 1 is to visualize the distribution of gluten and starch, which are components in bread dough, which is the measurement object 13, using fluorescent fingerprint imaging, and to compare with the stained image of the same sample, thereby fluorescent fluorescent imaging It is to verify the validity of the method.

(2. Experimental method)
(2-1. Acquisition of fluorescence fingerprint image data)
The dough at the pick-up stage and pure gluten and starch were frozen and a slice sample was prepared using a frozen microtome, and thawed and dried on a glass slide. The fluorescence image photographing device 10 used for obtaining the fluorescence fingerprint image data includes a spectral illumination device 11 and a spectral photographing device 12. In the spectral illumination device 11, the thin sample is irradiated with excitation light having a single wavelength from the xenon light source (MAX-302 (trade name), Asahi Spectroscopy (company name)) of the light source 110 that has passed through the bandpass filter of the spectral device 112. did. In the spectroscopic imaging device 12, the band-pass filter of the spectroscopic device 122 extracts only a fluorescent image of a specific wavelength from the sample image taken from the objective lens (M Plan Apo (trade name), Mitsutoyo (company name)) as the optical unit 120. Was taken with a cooled CCD camera (ORCA-ER-1394 (trade name), Hamamatsu Photonics (company name)) of the image capturing device 124. By capturing the same sample while changing the band-pass filters on the light source side and the camera side, fluorescence fingerprint image data that can reconstruct the fluorescence fingerprint data of each pixel was acquired by the processing of the fluorescence fingerprint information acquisition unit 24-1. .

(2-2. Visualization by cosine similarity)
By the processing of the similarity index calculation unit 24-2, for each pixel of the bread dough image, gluten and starch were discriminated by comparing the similarity between the fluorescence fingerprint of the image and the fluorescence fingerprint of pure gluten or starch. The cosine similarity is adopted as the similarity. Further, by the processing of the color determination unit 24-3 and the component distribution visualization unit 24-4, a color proportional to the value of the cosine similarity with gluten and starch is mapped to each pixel, whereby the abundance of each component is visually recognized. A pseudo image (for example, the images of FIGS. 11 to 16 described above) was constructed.

(2-3. Dyeing of sample and acquisition of stained image)
After obtaining an image and / or a composite image (for example, the composite image of FIG. 17 described above) as fluorescent fingerprint image data by processing of the component distribution visualization unit 24-4 and / or the composite image creation unit 24-5, the same sample is obtained. By visualizing both components with rhodamine B and FITC double staining to obtain a synthetic stained image (see FIG. 23 described later), for example, by comparing it with a synthetic image of the aforementioned pseudo image (see FIG. 17), The component distribution analysis method (fluorescence fingerprint imaging method) of the invention was verified.

(3. Experimental results)
FIG. 17 shows a composite image of bread dough visualized using the composite image creation unit 24-5 of this component distribution analysis method, and FIG. 23 shows a composite stained image of the same sample. 17 and 23, the red area indicated by (i) represents gluten, and the green area indicated by (ii) represents starch. Since the distribution of gluten and starch shown in both the images of FIG. 17 and FIG. 23 match, the validity of this component distribution analysis method was shown.

  Here, referring to FIGS. 21 to 23 and FIG. 17 described above, the validity of the composite image (for example, the composite image shown in FIG. 17 described above) created by the process of the composite image creation unit 24-5 is verified. The results will be described in more detail. Here, FIG. 21 is a diagram showing an example of a stained image of gluten in which gluten is stained red with rhodamine B. Moreover, FIG. 22 is a figure which shows an example of the dyeing | staining image of the starch which dye | stained the starch green by FITC. FIG. 23 is a diagram showing an example of a synthetic stained image obtained by synthesizing the gluten stained image of FIG. 21 and the starch stained image of FIG.

  Specifically, the stained images in FIG. 21 and FIG. 22 are generally used for visualization of gluten / starch using the measurement object 13 that has completed the fluorescence fingerprint measurement by the processing of the component distribution visualization unit 24-4. It is a stained image imaged by double staining using Rhodamine B and FITC. As an example, the observation conditions of the stained image of FIG. 21 in which gluten is dyed red using rhodamine B are: observation wavelength: excitation wavelength 540 nm, fluorescence wavelength: 605 nm. In addition, the observation conditions of the stained image in FIG. 22 in which starch is stained green using FITC are: observation wavelength: excitation wavelength 470 nm, fluorescence wavelength: 535 nm.

  Further, the synthetic stained image in FIG. 23 is obtained by using rhodamine B to verify the validity of the synthetic image (for example, the synthetic image in FIG. 17) acquired by the processing of the synthetic image creation unit 24-5. 21 and the stained image of FIG. 22 in which starch is stained green using FITC. 23 includes a red portion corresponding to gluten shown in (i) of FIG. 21 and a green portion corresponding to starch shown in (ii) of FIG.

  Then, when the synthesized image of FIG. 17 and the synthesized stained image of FIG. 23 are compared, the distribution of the red portion corresponding to gluten indicated by (i) of the synthesized image of FIG. The distribution of the red part corresponding to gluten indicated by i) is similar. Similarly, the distribution of the green part corresponding to the starch indicated by (ii) of the above-described composite image of FIG. 17 is similar to the distribution of the green part corresponding to the starch indicated by the composite dyed image (ii) of FIG. Yes. This means that the composite image in FIG. 17 appropriately visualizes the actual component distribution.

  Here, with reference to FIG. 24 and FIG. 25, the composite image created by the processing of the composite image creation unit 24-5 (for example, the composite image of FIG. 17) and the actual staining operation are taken. A complementary relationship between gluten and starch distribution will be described for a synthetic dyed image obtained by synthesizing the dyed image (for example, the synthetic dyed image of FIG. 23). Here, FIG. 24 is a diagram illustrating a complementary relationship between the cosine similarity with starch and the cosine similarity with gluten for the composite image in FIG. 17. FIG. 25 is a diagram showing a complementary relationship between the fluorescence intensity of starch and the fluorescence intensity of gluten in the synthetic stained image of FIG.

  Specifically, as shown in FIG. 24, values obtained by standardizing (ie, changing the scale) the cosine similarity with starch and the gluten are plotted, and these plot groups are aligned. It became a state. This is because the plot groups are arranged in a straight line, so there is a complementary relationship between the cosine similarity value with starch and the cosine similarity with gluten, and the composite image of FIG. 17 shows the component distribution of gluten and starch. Is expressed quantitatively. That is, as described above, in the composite image in FIG. 17, the distribution of gluten in the red part shown in (i) of FIG. 17 and the distribution of starch in the green part shown in (ii) of FIG. Is shown. That is, FIG. 17 shows that the portion with much gluten (that is, the red portion shown in FIG. 17 (i)) has less starch and the portion with much starch (that is, the green portion shown in FIG. 17 (ii)). Since it shows that there is little gluten, it shows that the state of actual distribution is appropriately reflected.

  On the other hand, as shown in FIG. 25, when the fluorescence intensity at 535 nm measured for FITC-stained starch and the fluorescence intensity at 605 nm measured for rhodamine B-stained gluten were plotted, the plot groups were not aligned on a straight line. Became dispersed. This means that there is no complementary relationship between the fluorescence intensity of starch and the fluorescence intensity of gluten (that is, there is no quantitative property) because the plot groups are not aligned and dispersed.

[Example 2]
Subsequently, in Example 2, the case where the measurement object 13 is bread dough is taken as an example, and the distribution of components (eg, gluten, starch, etc.) contained in the measurement object 13 is determined in order to determine the optimum degree of mixing. An example of quantification and calculation of the ratio of bubbles contained in the measurement object 13 (bubble area ratio) will be described.

  Specifically, in the present embodiment, the composite image created by the processing of the composite image creation unit 24-5 is divided into squares of a predetermined size, and the component in each square of the image or the composite image is divided. The component distribution was quantified by calculating the standard deviation of the ratio.

  Here, an example of quantifying the component distribution performed on the composite image created by the process of the composite image creation unit 24-5 will be described with reference to FIGS. In the present embodiment, an example will be described in which the component distribution is quantified for a composite image (a composite image shown in FIGS. 26 to 28 described later). The component distribution quantification process is performed by the component distribution visualization unit 24-4. The same processing can be performed on images acquired by the above processing (for example, the images shown in FIGS. 11 and 12, FIG. 15, and FIG. 16).

  Here, FIG. 26 is a diagram illustrating an example of a composite image of bread dough at the stage of insufficient mixing. FIG. 27 is a diagram showing an example of a composite image of bread dough at the optimum mixing stage. FIG. 28 is a diagram showing an example of a composite image of bread dough at an excessive mixing stage. FIG. 29 is a diagram illustrating an example of a component distribution quantification process when the component distribution is not uniform. FIG. 30 is a diagram illustrating another example of the component distribution quantification process when the component distribution is non-uniform. FIG. 31 is a diagram illustrating an example of a component distribution quantification process when the component distribution is uniform. FIG. 32 is a diagram illustrating another example of the component distribution quantification process when the component distribution is uniform. FIG. 33 is a diagram illustrating an example of the result of quantifying the component distribution.

  In this example, a model sample obtained by simplifying bread dough was used. Mixing was carried out according to the amount and time shown below to prepare a mixing-deficient fabric, an optimal mixing fabric, and an excessive mixing fabric, and 6 samples were measured at each stage.

Specifically, for example, a bread dough model sample was produced according to the following conditions.
・ Strong flour (Nisshin Seifun Camellia (trade name)) 3000 g
・ 2040 g of water (68% water against 100% strong powder)

From this mixture of strong powder and water, a dough of three mixing stages was prepared using a 20 L mixer.
・ Mixing short fabric Low speed 1 minute ・ Optimal mixing fabric Low speed 1 minute + Low speed 1 minute + Medium speed 4 minutes + High speed 2 minutes 30 seconds ・ Mixing excess fabric Low speed 1 minute + Low speed 1 minute + Medium speed 4 minutes + High speed 2 minutes 30 Second + high speed 7 minutes

  In addition, in order to make pure gluten and starch, the dough that had been crushed in the above-mentioned optimal mixing stage was mixed in water and divided into starch that flows with water and gluten that remains as a lump. In this embodiment, the fluorescence fingerprint information obtained from the starch and gluten is used as teacher data when analyzing the bread dough data (that is, the fluorescence fingerprint information of the target component obtained in advance that functions as a control). it can.

  For example, in the present embodiment, a composite image (for example, shown in FIGS. 26 to 28) of bread dough in three stages (for example, an insufficient mixing stage, an optimal mixing stage, an excessive mixing stage, etc.) by the process of the composite image creating unit 24-5. (Composite image) is assumed. As shown in FIG. 26, in the composite image of the bread dough at the stage of insufficient mixing, the degree of mixing of gluten corresponding to the red part of (i) and starch corresponding to the green part of (ii) is not uniform. Moreover, as shown in FIG. 27, in the composite image of the bread dough at the optimum mixing stage, the mixing degree of gluten corresponding to the red part of (iii) and starch corresponding to the green part of (iv) is uniform. Here, the portion (v) in FIG. 27 shows bubbles. As shown in FIG. 28, in the composite image of the dough at the excessive mixing stage, as in FIG. 27, the mixture of gluten corresponding to the red part of (vi) and starch corresponding to the green part of (vii) The condition is uniform, but includes larger bubbles (portions indicated by (viii) in FIG. 28) than bubbles indicated by the portion (v) in FIG. This shows a state where gluten and the like in the bread dough are cut off due to excessive mixing and the physical properties are changed. That is, there is a difference in distribution uniformity between the composite image in FIG. 26 and the composite image in FIG. 27, and there is a difference in the amount of bubbles between the composite image in FIG. 27 and the composite image in FIG. It is done.

  Specifically, in order to quantify the difference in the uniformity of the component distribution, as shown in FIG. 29, for example, the composite image of the dough at the stage of insufficient mixing shown in FIG. i) and (ii) is a 10 × 10 pixel small square). Further, as shown in FIG. 30, as a square having a predetermined size, for example, it may be divided into large squares of 40 × 40 pixels indicated by (vi) and (vii). Here, in FIGS. 29 and 30, (iii) indicates gluten, (iv) indicates starch, and (v) indicates bubbles. That is, in FIG. 29, since the ratio of components (gluten, starch, etc.) varies depending on the location (for example, for each small square indicated by (i) and (ii)), the dispersion of components such as gluten and starch is It is big. On the other hand, FIG. 30 shows that if the square is largely divided like the squares indicated by (vi) and (vii), the difference depending on the location is reduced.

  Further, as shown in FIG. 31, for example, a composite image of the bread dough at the optimum mixing stage in FIG. 27 and the bread dough at the excessive mixing stage in FIG. 28 is shown by squares of a predetermined size (for example, (i) and (ii)). Divided into small squares). Further, as shown in FIG. 32, as a square having a predetermined size, for example, it may be divided into large squares indicated by (vi) and (vii). Here, in FIGS. 31 and 32, (iii) indicates gluten, (iv) indicates starch, and (v) indicates bubbles. That is, FIG. 31 shows a state in which the ratio of components (gluten, starch, etc.) is almost uniform even in different places (for example, the small squares shown in (i) and (ii)). It shows that the dispersion of ingredients such as starch and starch is small. Further, FIG. 32 shows that the difference depending on the location becomes smaller if the square is divided into large portions, such as the squares indicated by (vi) and (vii).

Subsequently, for example, the standard deviation of the ratio R of the components in each square of the composite image of FIGS. 26 to 28 was calculated, and the component distribution was quantified. Here, the ratio R was calculated according to the following equation. In this example, as an example, component A was starch, component B was gluten, and the similarity index was cosine similarity.
Ratio R = similarity index of Σcomponent A / similarity index of Σcomponent B

  Here, an example of the quantification result of the component distribution will be described with reference to FIG. As shown in FIG. 33, the component distribution quantified in the present embodiment is represented by a standard deviation value for each side length (number of pixels) of the square. Here, in FIG. 33, ◯ indicates the value of the standard deviation calculated for the composite image of bread dough at the stage of insufficient mixing in FIG. The crosses indicate standard deviation values calculated for the composite image of bread dough at the optimum mixing stage in FIG. The + mark indicates the value of the standard deviation calculated for the composite image of the bread dough at the excessive mixing stage in FIG. That is, FIG. 33 shows that the bread dough components (gluten, starch, etc.) are largely dispersed in the mixing insufficient stage indicated by the circles, and the components of the bread dough in the optimal mixing stage indicated by the x marks and the excessive mixing stage indicated by the + signs. Since the dispersion of the ingredients of the bread dough is small, the ingredient distribution is expressed quantitatively.

  Subsequently, in this example, the bubble area ratio was calculated in order to quantitatively grasp the bubbles contained in the measurement object 13.

Specifically, in the present embodiment, the bubble area ratio indicating the bubble ratio in the image acquired by the processing of the component distribution visualization unit 24-4 or the composite image created by the processing of the composite image creation unit 24-5 The calculation was performed based on the total number of pixels (number of pixels) of the image or the composite image and the number of pixels including bubbles. Here, the bubble area ratio was calculated according to the following equation.
Bubble area ratio = number of pixels of bubbles / total number of pixels

  Subsequently, in this example, the optimum degree of mixing of the measurement object 13 was determined based on the component distribution quantified by the component distribution quantification process and the bubble area ratio calculated by the bubble area ratio calculation process. . The details are shown below.

  Here, with reference to FIG. 34 and FIG. 35, an example of the optimal mixing degree determination process in the present embodiment will be described. Here, FIG. 34 is a diagram illustrating an example of the uniformity of component distribution in each mixing stage. FIG. 35 is a diagram showing an example of the bubble area ratio in each mixing stage.

  Specifically, in the present embodiment, based on the component distribution quantified by the component distribution quantification process (see FIG. 33), each mixing stage as shown in FIG. 34 (for example, a mixing insufficient stage, an optimal mixing stage) The coefficient of variation (standard deviation / average) indicating the uniformity of distribution in the mixing excess stage, etc.) was calculated, and the optimum mixing degree of the measurement object was determined. Here, in FIG. 34, when the insufficient mixing stage and the optimal mixing stage (or the excessive mixing stage) are compared, the uniformity of the component distribution is significantly increased (p <0.01). On the other hand, there was no significant difference in the uniformity of component distribution between the optimal mixing stage and the overmixing stage. In this way, in the optimum mixing degree determination process in the present embodiment, the distribution of gluten and starch becomes uniform when shifting from the insufficient mixing stage to the optimum mixing stage, so the bread dough at the optimum mixing stage (see FIG. 27). Is determined to be in the state of optimal mixing.

  Furthermore, in this embodiment, based on the bubble area ratio calculated by the bubble area ratio calculation process, the measurement at each mixing stage (for example, the mixing insufficient stage, the optimum mixing stage, the mixing excessive stage, etc.) as shown in FIG. The optimum degree of mixing of the object was determined. Here, in FIG. 35, when the mixing insufficient stage (or the optimal mixing stage) and the excessive mixing stage were compared, the bubble area ratio was significantly increased (p <0.05). On the other hand, there was no significant difference in the bubble area ratio between the insufficient mixing stage and the optimum mixing stage. Thus, in the optimum mixing degree determination process in the present embodiment, the bubble area ratio increases when shifting from the optimum mixing stage to the excessive mixing stage, so that the bread dough (see FIG. 27) in the optimum mixing stage is optimally mixed. Judged to be in the state of degree.

  As described above, it has been difficult to quantitatively grasp the optimum mixing state with a conventional stained image or the like. However, according to the present invention, the component distribution can be visualized quantitatively, and the optimum mixing stage can be visualized. It becomes easy to grasp and can improve the product quality stability.

[Other embodiments]
Although the embodiments of the present invention have been described so far, the present invention is not limited to the above-described embodiments, but can be applied to various different embodiments within the scope of the technical idea described in the claims. It may be implemented.

  In addition, among the processes described in the embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method.

  In addition, unless otherwise specified, the processing procedures, control procedures, specific names, information including registration data for each processing, parameters such as search conditions, screen examples, and database configurations shown in the above documents and drawings Can be changed arbitrarily.

  In addition, regarding the component distribution analysis system, each illustrated component is functionally conceptual and does not necessarily need to be physically configured as illustrated.

  For example, the processing functions of each component of the component distribution analyzer 20, particularly the processing functions performed by the calculation processing unit 24, are all or any part of the processing functions by a CPU (Central Processing Unit) and the CPU. It may be realized by a program to be interpreted and executed as hardware by wired logic. The program is recorded on a recording medium to be described later, and is mechanically read by the component distribution analyzer 20 as necessary. That is, a memory 21 such as a ROM or an HD stores a computer program for giving instructions to the CPU in cooperation with an OS (Operating System) and performing various processes. This computer program is executed by being loaded into the RAM, and constitutes a control unit in cooperation with the CPU.

  The computer program may be stored in an application program server connected to the component distribution analyzer 20 via an arbitrary network, and may be downloaded in whole or in part as necessary. It is.

  In addition, the program according to the present invention may be stored in a computer-readable recording medium, and may be configured as a program product. Here, the “recording medium” is any memory card, USB memory, SD card, flexible disk, magneto-optical disk, ROM, EPROM, EEPROM, CD-ROM, MO, DVD, Blu-ray Disc, etc. Of “portable physical media”.

  The “program” is a data processing method described in an arbitrary language or description method, and may be in any format such as source code or binary code. The “program” is not necessarily limited to a single configuration, but is distributed in the form of a plurality of modules and libraries, or in cooperation with a separate program represented by an OS (Operating System). Including those that achieve the function. Note that a well-known configuration and procedure can be used for a specific configuration for reading a recording medium, a reading procedure, an installation procedure after reading, and the like in each device described in the embodiment.

  Various databases and the like stored in the memory 21 are storage means such as a memory device such as a RAM and a ROM, a fixed disk device such as a hard disk, a flexible disk and an optical disk, and various programs and programs used for providing websites. Stores tables, databases, files for web pages, etc.

  The component distribution analyzer 20 may be configured as an information processing apparatus such as a known personal computer or workstation, or may be configured by connecting any peripheral device to the information processing apparatus. The component distribution analyzer 20 may be realized by installing software (including programs, data, and the like) that causes the information processing apparatus to realize the method of the present invention.

  Furthermore, the specific form of distribution / integration of the devices is not limited to that shown in the figure, and all or a part of them may be functional or physical in arbitrary units according to various additions or according to functional loads. Can be distributed and integrated. That is, the above-described embodiments may be arbitrarily combined and may be selectively implemented.

  As described above in detail, according to the present invention, in the component distribution analysis for visualizing the distribution of the target component from the obtained fluorescent fingerprint, the target component can be further identified, and the component Since a component distribution analysis method, a component distribution analysis device, and a program can be provided, it is possible to suitably use for various applications such as analysis of the internal structure of food or the like.

  In addition, the present invention is a very versatile technique that can be used from a basic research area in which the internal structure of various substances is examined to an applied field such as inspection at a food manufacturing site. The scale of measurement ranges from the micro scale observed with an optical microscope to the macro scale that can be imaged with a normal camera, and is not limited to the food field, but can be expected to have a wide range of applications from agriculture to living body / medical field. Since non-destructive measurement is possible in the present invention, if the measurement time can be shortened by narrowing down the measurement conditions, it can also be applied to 100% inspection in a factory or real-time monitoring.

DESCRIPTION OF SYMBOLS 10 Fluorescence image imaging device 11 Spectral illumination device
110 Light source
112 Spectrometer
114 Excitation wavelength adjustment unit
116 optical fiber
118 Excitation light irradiation unit 12 Spectroscopic imaging device
120 Optics
122 Spectrometer
124 Image Capture Device 13 Measurement Object 20 Component Distribution Analysis Device 21 Memory 22 Keyboard / Mouse 23 Control Unit 24 Calculation Processing Unit
24-1 Fluorescent fingerprint information acquisition unit
24-2 Similarity index calculator
24-3 Color decision unit
24-4 Component distribution visualization unit
24-5 Composite Image Creation Unit 30 Display

Claims (15)

A component distribution analysis method for analyzing a measurement object,
Measuring the fluorescence intensity at a plurality of measurement locations 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, Fluorescent fingerprint information acquisition process for acquiring fluorescent fingerprint information of each measurement location on the object,
A similarity index is calculated by comparing the fluorescent fingerprint information of the measurement object acquired for each measurement location in the fluorescent fingerprint information acquisition step with fluorescent fingerprint information of a target component acquired in advance. Similarity index calculation process,
A color determination step for determining a color to be colored at each measurement location based on the similarity index calculated in the similarity index calculation step; and
Color distribution of the color to be colored determined in the color determination step to each measurement location, and a component distribution visualization step of acquiring an image in which the component included in the measurement object is visualized,
A component distribution analysis method comprising: The image acquired in the component distribution visualization step and another image obtained by visualizing a component different from the component of the image are synthesized to simultaneously combine a plurality of components included in the measurement object. A composite image creation process for creating a visualized composite image;
The component distribution analysis method according to claim 1, further comprising: The similarity index is
Cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, Morisita similarity index, standard Euclidean distance similarity, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity The component distribution analysis method according to claim 1, wherein the component distribution analysis method is at least one of a degree, a Jackard coefficient similarity, a Dice coefficient similarity, and a Simpson coefficient similarity.   The component distribution analysis method according to claim 1, wherein the component is a component derived from a living body.   The component distribution analysis method according to claim 4, wherein the biological component includes at least one selected from the group consisting of protein, starch, polysaccharides, minerals, and lipids. A component distribution analyzer for analyzing a component distribution of the measurement object, comprising: a spectral illumination device that irradiates the measurement object with a predetermined excitation wavelength; and a spectroscopic imaging device that observes the measurement object with a predetermined fluorescence wavelength Because
Measuring the fluorescence intensity at a plurality of measurement locations 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, A fluorescence fingerprint information acquisition unit for acquiring fluorescence fingerprint information of each measurement location on the object;
Similarity of calculating the similarity index by comparing the fluorescent fingerprint information of the measurement object acquired for each measurement location by the fluorescent fingerprint information acquisition unit with the fluorescent fingerprint information of the target component acquired in advance Degree index calculator,
A color determining unit that determines a color to be colored at each measurement location based on the similarity index calculated by the similarity index calculating unit; and
A component distribution visualization unit for color-mapping the color to be colored determined by the color determination unit at each measurement location, and obtaining an image that visualizes the component included in the measurement object;
A component distribution analysis apparatus comprising: By synthesizing the image acquired by the component distribution visualization unit and another image obtained by visualizing a component different from the component of the image, a plurality of components included in the measurement object are simultaneously visualized. A composite image creation unit for creating the composite image
The component distribution analyzer according to claim 6, further comprising: The similarity index is
Cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, Morisita similarity index, standard Euclidean distance similarity, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity The component distribution analyzer according to claim 6, wherein the component distribution analyzer is at least one of degree, Jackard coefficient similarity, Dice coefficient similarity, and Simpson coefficient similarity.   The component distribution analyzer according to any one of claims 6 to 8, wherein the component is a component derived from a living body.   The component distribution analyzer according to claim 9, wherein the biological component includes at least one selected from the group consisting of protein, starch, polysaccharides, minerals, and lipids. A component distribution analyzer for analyzing a component distribution of the measurement object, comprising: a spectral illumination device that irradiates the measurement object with a predetermined excitation wavelength; and a spectroscopic imaging device that observes the measurement object with a predetermined fluorescence wavelength A program for executing
In the component distribution analyzer,
Measuring the fluorescence intensity at a plurality of measurement locations 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, Fluorescent fingerprint information acquisition process for acquiring fluorescent fingerprint information of each measurement location on the object,
A similarity index is calculated by comparing the fluorescent fingerprint information of the measurement object acquired for each measurement location in the fluorescent fingerprint information acquisition step with fluorescent fingerprint information of a target component acquired in advance. Similarity index calculation process,
A color determination step for determining a color to be colored at each measurement location based on the similarity index calculated in the similarity index calculation step; and
Color distribution of the color to be colored determined in the color determination step to each measurement location, and a component distribution visualization step of acquiring an image in which the component included in the measurement object is visualized,
A program for running The image acquired in the component distribution visualization step and another image obtained by visualizing a component different from the component of the image are synthesized to simultaneously combine a plurality of components included in the measurement object. A composite image creation process for creating a visualized composite image;
The program according to claim 11, further comprising: The similarity index is
Cosine similarity, Pearson correlation coefficient, deviation pattern similarity, Euclidean distance similarity, Morisita similarity index, standard Euclidean distance similarity, Mahalanobis distance similarity, Manhattan distance similarity, Chebyshev distance similarity, Minkowski distance similarity 13. The program according to claim 11, wherein the program is at least one of degree, Jackard coefficient similarity, Dice coefficient similarity, and Simpson coefficient similarity.   The program according to claim 11, wherein the component is a biological component.   The program according to claim 14, wherein the biological component includes at least one selected from the group consisting of protein, starch, polysaccharides, minerals, and lipids.

Images