JPH10311792A - Method and equipment for measuring moisture of tea leaf - Google Patents

Abstract

PROBLEM TO BE SOLVED: To measure the moisture of teat leaf accurately by determining the Kubelka-Munk functional value from a measured diffuse reflection light of a teat leaf and applying the functional values to a working curve recurrence formula based on samples having known moisture content thereby determining a moisture content. SOLUTION: A teat leaf 5 is held a measuring sample 6 on the front face of an optical transparent plate 13. A detector 28 detects the light transmitted through each interference filter 24 and reflected on a standard reflector 27'. The standard reflector 27' is then retracted and the diffuse reflection light from the measuring sample 6 is measured for each wavelength and each detection value is outputted to a control section 3. The control section 3 determines a Kubelka-Munk functional value (F value) from a measured diffuse reflection light of a teat leaf having known moisture and produces a working curve recurrence formula representative of relationship between the F value and the moisture content. Similarly, an F value is determined from a measured diffuse reflection light of a teat leaf having unknown moisture and applied to a working curve recurrence formula thus determining the moisture content of a teat leaf.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a measuring apparatus and a measuring method for measuring the water content of tea leaves, that is, the water content.

[0002]

2. Description of the Related Art A green tea production process comprises a plurality of steps in which raw tea leaves are steamed and then drying and shaping operations are sequentially performed while applying rubbing pressure. The process is divided into the steps of fresh leaf reception, fresh leaf management, foliage, coarse kneading, kneading, medium kneading, fine kneading, and drying, depending on the degree of drying.

[0003] In controlling the above-mentioned tea making processes, a skilled worker observes the condition of the tea leaves to be processed and the operating condition of the machine in each process and adjusts the machine so as to obtain the best condition. It is generally done. on the other hand,
In recent years, the automation of tea-making machines and the scale of tea-making plants have been progressing, and efforts have been made to reduce labor such as the loading and unloading of tea leaves in each process and the transporting work to the next process. In the large-scale tea plant, the automatic control of the tea making process, which samples the tea leaves during the tea making process, detects the processing state of the tea leaves and controls the operation conditions of each of the following processes according to the detection result, It is taking place.

[0004] In the case where the above-mentioned process control by a skilled worker or the automatic control of the tea-making process performed in a large-scale tea-making plant is performed, in setting the process conditions in the tea-making process, the moisture content of the sample from the previous process is not considered. Accordingly, the process conditions after the next process are determined, and it is important to perform the water management and the measurement of the water state of the tea leaves during the process accurately and promptly.

Therefore, various devices for measuring the moisture content during the tea making process have been proposed, for example, a method for measuring the electrical conductivity of tea leaves (Japanese Patent Application Laid-Open No. 3-39041) and a method using microwaves (Japanese Patent Application Laid-Open No. Hei 2-39041). Japanese Unexamined Patent Publication (Kokai) No. -39853), and those using a high-frequency capacitance type moisture meter have been proposed. However, these tend to vary in measurement accuracy due to the influence of the packing density of the sample and the timing of extracting tea leaves. In addition, there is a problem that measurement in a high moisture region is difficult, and the cost is high and practical use has not been achieved.

On the other hand, measurement accuracy, quickness of measurement,
Proposals for controlling the tea-making process by applying moisture measurement by a near-infrared spectrometer to the moisture measurement in the tea-making process because it is effective in terms of cost have been proposed, for example, in JP-A 1-21
This is performed as described in JP-A-5241.

[0007]

However, the use of a near-infrared spectrometer has the following problems. That is, the relationship between the water content of the tea leaves and the output signal obtained by the near-infrared analyzer is a non-linear relationship as shown in FIG. Moreover, the change in the water content of the tea leaves is quite wide, such as 80-5% between the steps described above. For this reason, even if an attempt is made to predict the water content by a multiple regression equation in consideration of the absorbance (log (1 / R)) using diffuse reflection light at 1940 nm, which is a single wavelength of water attribution, it is linear within the above wide range. It is only possible to predict the moisture value for a very narrow range in which the relationship can be obtained, and if a single multiple regression equation is used to predict the moisture value over the entire process, the error will increase and accurate moisture At present it is not possible to measure the value. Therefore, a calibration curve was set for each of the steps of leafing, coarse kneading, kneading, medium kneading, fine kneading, and drying, and based on the output signal from the near-infrared analyzer in each step, these calibration curves were used for tea leaves. It is conceivable to use differently according to the water content. However, in such a method, not only is the labor for preparing a calibration curve corresponding to the moisture content of the tea leaves in each step enormous, but also the cost for preparing the calibration curve may increase. on the other hand,
The near-infrared ray used for moisture prediction is not only 1940 nm which is a single wavelength attributed to moisture, but a few wavelengths are prepared as correction wavelengths other than the above-mentioned wavelengths, or a plurality of types according to the content state of moisture. Wavelength is used. For example, at high moisture, the wavelength is 1450 nm or even shorter.
70 nm, 760 mn or a wavelength close to this is used,
At low moisture, a wavelength of 1940 nm or a wavelength close to 1940 nm is used for the purpose of increasing the absorption sensitivity. As described above, a plurality of wavelengths are prepared in order to cope with a wide range of water content, and those wavelengths are selected at the time of moisture estimation.
However, when a plurality of types of wavelengths are prepared, it is necessary to set a calibration curve for each wavelength and to prepare a filter for each wavelength in the optical system and to perform maintenance such as a switching operation. Moreover, the reason for preparing such a plurality of wavelengths is that, besides simply targeting the moisture content region, the absorbance varies greatly due to the variation in the particle size of the sample used for measuring moisture such as tea leaves. In some cases. For this reason, there is a problem that the labor and structure required for moisture estimation increase.

A first object of the present invention is to provide an apparatus for measuring the water content of a tea leaf, which can accurately measure the water content without complicating the structure, that is, can predict the water content, in view of the above-mentioned problems in the conventional measurement of the water content of the tea leaf. Is to do.

A second object of the present invention is to solve the above-mentioned problems in the conventional measurement of water content of tea leaves, and to unify the calibration curve formula to enable highly accurate prediction of water content, thereby enabling the water content of tea leaves to be estimated. An object of the present invention is to provide a moisture measurement method capable of simplifying the time and effort required for prediction.

[0010]

Means for Solving the Problems In order to achieve this object, the invention according to claim 1 is directed to a moisture measuring apparatus for measuring moisture of tea leaves in a tea making process using near infrared rays.
A measuring unit closed via an optical transparent plate with respect to the transport space in which the tea leaves move, and is installed in the transport space, operates when measuring the moisture of the tea leaves, and is disposed on the front surface of the optical transparent plate. A Kubelka-Munk function value (F-value) of the tea leaf is obtained using a Kubelka-Munk function from the sample supply means for supplying the tea leaf and the diffusely-reflected light of the tea leaf measured by the measuring unit. And a control unit for calculating the water content of the tea leaves for which the Kubelka-Munk function value (F value) has been determined by a calibration curve regression formula created using the control unit.

According to a second aspect of the present invention, in the tea leaf moisture measuring apparatus according to the first aspect, the near-infrared irradiation optical system in the measuring section is 1940 nm for one wavelength and 1940 nm and 2139 nm for two wavelengths. Are selectable.

The invention described in claim 3 is the first or second invention.
A moisture measuring method using the moisture measuring device for tea leaves according to the above, wherein the sample whose moisture is known is irradiated with near-infrared rays by an i-th filter, and diffusely reflected light (Ri) from the sample is measured. The Kubelka-Munk function value (F) is calculated from the diffuse reflection light (Ri) using the Kubelka-Munk function.
i) is determined by Fi = (1−Ri) ² / (2Ri), where R: diffuse reflectance, and a calibration curve regression equation using the above Kubelka-Munk function value (Fi) and known moisture, G (% ) = Ka + K ₁ · F ₁ + K ₂ · F ₂ ... + Kn · F
n, where ka is a constant term (bias value), Kn is a regression coefficient in the nth filter, and Fn is a Kubelka-Munk function value (F value) in the nth filter. Irradiate near-infrared light with a filter and diffusely reflect light (Ri ') from the above sample
Is measured, and a Kubelka-Munk function value (Fi ′) is obtained from the diffuse reflection light (Ri ′) using a Kubelka-Munk function.
Where Fi ′ = (1−Ri ′) ² / (2Ri ′) where R ′: diffuse reflectance, and the Kubelka-Munk function value (Fi ′)
Is applied to the calibration curve regression equation created for the above-described sample with known moisture to determine the moisture value of the sample with unknown moisture.

[0013]

According to the first and second aspects of the invention, the Kubelka-Munk function value (F-value) is arithmetically processed by the Kubelka-Munk function based on the diffusely-reflected light from the tea leaves measured by the measuring unit, thereby obtaining the reflected light and the water content. Is linearized. By applying the F value to the calibration curve regression equation, it is possible to use multiple types of wavelengths and interference filters without using
The moisture value of the tea leaf can be predicted using a single linear model relating to the relationship between the F value and the moisture content.

According to the third aspect of the present invention, the above-mentioned F value for a sample having a known water content is obtained by a Kubelka-Munk function, and a calibration curve regression equation is prepared using the F value and the known water content. The moisture value of the unknown moisture sample can be predicted simply by applying the F value obtained by the Kubelka-Munk function to the sample with unknown moisture to this calibration curve regression equation. This makes it possible to predict the moisture value of a sample whose moisture content is unknown over the entire moisture content range by a simple process using only a single calibration curve regression equation based on the Kubelka-Munk function.

[0015]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. FIG. 1 is a schematic diagram for explaining an embodiment of the invention described in claims 1 to 3. In FIG. 1, reference numeral 1 denotes
Fresh leaf receipt, fresh leaf management, steaming, leaf beating, coarse kneading, kneading, medium kneading,
A transport space for transporting tea leaves to pass through each of the tea making processes of fine rubbing and drying, reference numeral 2 denotes a measuring unit provided separately from the transport space 1, and reference numeral 3 denotes an arithmetic processing of a signal from the measuring unit. Reference numeral 4 denotes a control unit, and reference numeral 4 denotes a display unit that displays a calculation result in the control unit 3.

A transfer means such as a conveyor (not shown) is provided in the transfer space 1, and the tea leaves move in the transfer space 1 by the transfer means. An opening is formed between the transport space 1 and the measuring unit 2, and the opening is formed in the opening.
An optical transparent plate 13 is provided so as to close the opening. In addition, the moving plate 1 that slides in and out of the transport space 1 so as to cross the moving direction of the tea leaves in the transport space 1.
2 is provided, and this moving plate 12 is an optically transparent plate 1
3 constitutes a sample supply means for supplying a measurement sample such as tea leaves to the front surface.

In the measuring section 2 which is closed by being separated from the transport space 1 by the optical transparent plate 13, the light from the light source 21 is optically transparent via the lens 22, the chopper disk 23, the interference filter, and the aperture 24. An optical system is arranged to irradiate the plate 13. The measuring section 2 is provided with a standard reflection plate 17 that protrudes and retracts on the optical axis in front of the optically transparent plate 13 and a photodetector 28 that detects light reflected from a sample such as tea leaves. Have been.

The main part of the control unit 3 is constituted by a microcomputer which can be operated and controlled, and a detector 28 is provided on the input side via an I / O interface (not shown).
However, the display unit 4 is connected to the output side. The control unit 3 first targets a tea leaf as a sample having a known moisture content, and obtains a Kubelka-Munk function value (hereinafter, referred to as a “Kverka-Munk function”) from a measured value of diffuse reflection light detected from the tea leaf by the Kubelka-Munk function equation shown in Expression (1). F value)
Is required. In this case, the Kubelka-Munk function value (Fi) when the i-th filter is used for tea leaves with a known moisture content is obtained by the Kubelka-Munk function formula (1).

Fi = (1−Ri) ² / (2Ri) (1) where R: diffuse reflectance

This F value is obtained for each irradiation light of each wavelength, and is used for obtaining the water content.

Next, the control unit 3 performs a calculation process for obtaining the following calibration curve regression equation (2) from the obtained F value and obtaining the water content of the measurement sample. The calibration curve regression equation is a regression equation that obtains an F value for a sample with a known moisture using the above-described Kubelka-Munk function equation, and summarizes a correlation between the F value and the water content G (%). , And a multiple regression equation using F values at several wavelengths as explanatory variables.

G (%) = Ka + K ₁ · F ₁ + K ₂ · F ₂ +... + Kn · Fn (2) where Ka: constant term (bias value), Kn: regression coefficient in the nth filter, Fn: Kubelka-Munk function value (F value) in the n-th filter

The control unit 3 previously obtains (Ka, K ₁ , K ₂ ,..., Kn) which are the constants of the regression equation of the calibration curve, registers them in a memory, and stores the F value obtained for the measurement sample. Is substituted into the calibration curve regression equation (2) to obtain the water content G of the unknown measurement sample. The value of the moisture content obtained for the measurement sample is sent to the display unit 4 and displayed. In the measuring section 2, when only one wavelength is used, a wavelength of 1940 nm is used, and when two wavelengths are used, 1940 nm and 2139 nm are selected as an example.

Since the present embodiment has the above configuration,
The operation of the moisture measuring device and a moisture measuring method using the device will be described. At the time of moisture measurement, the movable plate 12 is manually or automatically operated and slides to a position 12 'indicated by a broken line, and a part of the tea leaves 5 is pressed against the front surface of the optical transparent plate 13 with a constant pressure, and is held as the measurement sample 6. I do. Prior to this, in the measuring section, the standard reflection plate 27 made of ceramic or the like is moved to a position 27 'shown by a broken line and set on the optical axis of the light from the light source 21. Light source 21
Are converted into parallel rays by the lens 22, are periodically divided by the chopper disk 23, and become monochromatic lights by the interference filter 24. The interference filter 24 is provided with a plurality of openings at predetermined intervals at a position equidistant from the center of the disk, and a filter element that transmits near infrared rays of a predetermined wavelength is attached to the openings. The filter element is configured to rotate so that one of the filter elements is positioned on the optical axis of the lens 22 when the filter element is stopped.

The light transmitted through the interference filter 24 is first applied to a standard reflection plate 27, and the reflected light is detected by a detector 28 to measure the reference reflected light. This measurement is performed for all the filters by automatically switching the interference filters 24 sequentially. Next, the standard reflection plate 27 is retracted from the front surface of the optical transparent plate 13, and the measurement sample 6 sandwiched between the front surfaces of the optical transparent plate 13 is similarly irradiated with light via the optical transparent plate 13. At each wavelength, diffusely reflected light from the measurement sample is measured. The measured detected values of the reflected light are converted into digital signals and output to the control unit 3. In the above, the case of the pre-spectroscopy by the interference filter has been described. However, the position of the spectroscopy may be the post-spectroscopy, and the spectroscopy method may be a grating.

The control unit 3 first obtains the above F value from the detected diffusely reflected light for a tea leaf as a sample with known moisture. In this case, the F value (Fi) using the i-th filter with respect to the sample whose water content is known is obtained by the above Kubelka-Munk function formula (1).

Next, the control unit 3 creates a calibration curve regression equation based on the obtained F value for the sample with known moisture. In this case, the above-described calibration curve regression equation (2) is used as the calibration curve regression equation. On the other hand, when the calibration curve regression equation is created for a sample with a known moisture, a process for obtaining the moisture content of a sample with an unknown moisture is then performed. That is, when tea leaves with unknown moisture are used as measurement samples, diffuse reflection light is measured in the same manner as in the above processing, and the measured value is applied to the Kubelka-Munk function formula (1) to obtain the F value of tea leaves with unknown moisture. (For the sake of convenience, Fi ′ is used to distinguish F values from samples with known moisture.) When the F value (Fi ′) of the tea leaf with unknown moisture is obtained, the F value (Fi ′) is applied to a previously created calibration curve regression equation (2) to obtain the moisture content (G ′) of the tea leaf with unknown moisture. ). Next, the control unit 3 outputs the water content (G) obtained by the calibration curve regression equation (2) to the display unit 4 and causes the display unit 4 to display it.

According to the above-described embodiment, the Kubelka-Munk function value (F-value) is obtained by using the Kubelka-Munk function, and as shown in FIG. 2, the F-value obtained from the diffuse reflection light and the water content are obtained. Since the relationship is linearly modeled and the moisture content can be predicted by the calibration curve regression equation based on the linear model, even if the moisture content region is wide, simply preparing a single linear regression model will reduce the unknown moisture content. The water content of tea leaves can be accurately determined. Moreover, the present inventor conducted a comparison between a predicted value based on the moisture content obtained in the above example and the moisture content of tea leaves actually used as a sample. As shown in FIG. A non-linear state was not obtained over the moisture content region, and there was no variation between the predicted value and the actually measured value. In other words, the result was that high accuracy could be maintained. The moisture measuring device shown in the above embodiment can be applied to the tea making process either online or at-line, and can predict the moisture of tea leaves in each process.

[0029]

According to the first and second aspects of the present invention,
The relationship between the reflected light and the water content is linearized by calculating the Kubelka-Munk function value (F-value) using the Kubelka-Munk function based on the diffusely reflected light from the tea leaves measured by the measuring unit. Predicting tea leaf moisture using a single linear model of the relationship between F value and moisture content without using multiple wavelengths or interference filters by applying the values to a calibration curve regression equation Becomes possible.

According to the third aspect of the present invention, the Kubelka-Munk function value (F value) for a sample with a known moisture content
By the Kubelka-Munk function, and furthermore, this F
A calibration curve regression equation is created using the values and the known moisture,
By applying the F value obtained based on the Kubelka-Munk function to a sample with unknown moisture to the calibration curve regression equation, the moisture value of the sample with unknown moisture can be predicted. This makes it possible to predict the moisture value of a sample whose moisture content is unknown over the entire moisture content region by a simple process using only a single calibration curve regression equation based on the Kubelka-Munk function.

[Brief description of the drawings]

FIG. 1 is a schematic diagram for explaining an embodiment of the invention described in claims 1 to 3;

FIG. 2 is a diagram for explaining a calibration curve obtained by the embodiment shown in FIG. 1;

FIG. 3 is a diagram for explaining an error between a predicted value and a measured value relating to the moisture content of tea leaves according to the embodiment shown in FIG. 1;

FIG. 4 is a diagram for explaining a calibration curve obtained by a conventional moisture measurement method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Sample placement part 11 Conveyor 12 Moving plate 13 Optically transparent plate 2 Measuring part 3 Control part 4 Display part 5 Tea leaf 28 Detector

Claims (3)

[Claims] 1. A moisture measuring device for measuring the moisture of tea leaves in a tea making process using near-infrared rays, comprising: a measuring section closed via a transparent optical plate with respect to a transport space in which the tea leaves move; It is installed in, and operates at the time of moisture measurement of the tea leaves, sample supply means for supplying the tea leaves to the front surface of the optical transparent plate, from the diffuse reflection light of the tea leaves measured by the measurement unit, Kubelka-Munk function Is used to determine the Kubelka-Munk function value (F value) of the tea leaf, and the water content of the tea leaf from which the Kubelka-Munk function value (F value) is determined by a calibration curve regression equation created using a sample whose moisture content is known in advance is calculated. And a control unit for determining the water content of the tea leaves. 2. The tea leaf moisture measuring apparatus according to claim 1, wherein the near-infrared irradiating optical system in the measuring section is 1940 nm for one wavelength and 1940 nm and 213 for two wavelengths.
An apparatus for measuring water content of tea leaves, wherein 9 nm can be selected. 3. A moisture measuring method using the moisture measuring device for tea leaves according to claim 1 or 2, wherein the sample whose moisture is known is irradiated with near-infrared rays by an i-th filter, and the sample is diffused from the sample. The reflected light (Ri) is measured, and a Kubelka-Munk function value (Fi) is calculated from the diffuse reflected light (Ri) by using a Kubelka-Munk function, where Fi = (1−Ri) ² / (2Ri) where R: diffuse reflectance G (%) = Ka + K ₁ · F ₁ + K ₂ · F ₂ ... + Kn · F which is a calibration curve regression equation using the above Kubelka-Munk function value (Fi) and known moisture.
n, where ka is a constant term (bias value), Kn is a regression coefficient in the nth filter, and Fn is a Kubelka-Munk function value in the nth filter. And diffusely reflected light (Ri ′) from the sample is measured. From the diffusely reflected light (Ri ′), a Kubelka-Munk function value (Fi ′) is calculated using the Kubelka-Munk function, and Fi ′ = (1−Ri ′). ² / (2Ri ′) where R ′: Diffuse reflectance, and the above Kubelka-Munk function value (Fi ′) is applied to a calibration curve regression equation created for the above-mentioned sample with known moisture, and the unknown moisture sample is determined. A method for measuring the moisture content of tea leaves, which comprises determining a moisture value.