CN112461820A - Tea amino acid content determination method based on colorimetric capsule image recognition - Google Patents
Tea amino acid content determination method based on colorimetric capsule image recognition Download PDFInfo
- Publication number
- CN112461820A CN112461820A CN202011241593.2A CN202011241593A CN112461820A CN 112461820 A CN112461820 A CN 112461820A CN 202011241593 A CN202011241593 A CN 202011241593A CN 112461820 A CN112461820 A CN 112461820A
- Authority
- CN
- China
- Prior art keywords
- amino acid
- capsule
- sample
- solution
- colorimetric
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/75—Systems in which material is subjected to a chemical reaction, the progress or the result of the reaction being investigated
- G01N21/77—Systems in which material is subjected to a chemical reaction, the progress or the result of the reaction being investigated by observing the effect on a chemical indicator
- G01N21/78—Systems in which material is subjected to a chemical reaction, the progress or the result of the reaction being investigated by observing the effect on a chemical indicator producing a change of colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Chemical & Material Sciences (AREA)
- Plasma & Fusion (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Health & Medical Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By The Use Of Chemical Reactions (AREA)
Abstract
The invention discloses a tea amino acid content determination method based on colorimetric capsule image recognition, and belongs to the field of food detection. The method comprises the following steps: (1) preparing a color reaction reagent phosphate buffer solution and a ninhydrin solution, and respectively freezing and solidifying the color reaction reagent phosphate buffer solution and the ninhydrin solution into a plurality of unit blocks with fixed quantity to prepare a capsule 1 and a capsule 2; (2) the tea sample is hot water extract; (3) melting the capsules 1 and 2, and reacting with the tea extract to obtain a sample color developing solution; (4) photographing and collecting a sample color development image, and extracting image RGB data; (5) inputting the image data into a standard concentration model, and outputting an amino acid predicted concentration value in an operating mode; the standard concentration model is a concentration-image mathematical model established by taking glutamic acid as amino acid standard solution. The invention establishes a model with standard concentration corresponding to the image, predicts the amino acid content through image recognition, and can quickly and accurately obtain the free amino acid content in the tea.
Description
Technical Field
The invention belongs to the field of food detection, and relates to a method for rapidly detecting the content of amino acids in tea leaves based on colorimetric capsule image recognition.
Background
Tea has become increasingly popular as a healthy drink for drinking and consumption in daily life. However, the tea leaves are various in category, and the tea leaves in the market are mixed, so that great difficulty is brought to daily shopping of consumers.
The quality judgment of tea leaves is mainly based on manual review, but general consumers cannot reach the level of professional review. The laboratory is mostly used for analyzing the internal quality of the tea by detecting physicochemical components, but the laboratory detection has the problems of large instrument volume, complex operation and the like, and is difficult to popularize in daily life.
The amino acid is one of the main chemical components in the tea, has close relationship with the aroma and the taste of the tea and plays an important role in the formation of the quality of the tea. The method for measuring the free amino acid in the tea leaves, which is recorded by the current national standard method (measurement of the total amount of the free amino acid in the tea leaves from GBT8314 and 2013), comprises the following steps: and (3) carrying out color comparison by using ninhydrin, measuring by using an ultraviolet spectrophotometer to obtain absorbance, and calculating to obtain the amino acid content value by referring to a standard curve. The principle is that alpha-amino acid and ninhydrin are heated together under the condition of pH8.0 to form purple complex, and the content of the complex is measured by spectrophotometry at specific wavelength.
The ultraviolet spectrophotometer which is a detection instrument required by the method is laboratory equipment, has large volume and high cost, needs instrument operation training and is not suitable for daily use. The buffer solution and the color developing agent need to be prepared in advance before each experiment, and the operation is complicated. In each detection, a standard solution needs to be prepared, and a standard curve is established. The quantity of samples for single detection is limited, and batch detection cannot be carried out. The instrument detects the obtained absorbance data, and further calculation is needed to obtain the amino acid content.
Therefore, there is a need to improve the existing method, simplify the detection steps, and achieve the purpose of quickly and accurately determining the content of free amino acids in tea leaves.
Disclosure of Invention
The invention aims to provide a method capable of quickly and accurately measuring the content of free amino acids in tea leaves, and aims to solve the problems that the existing method is complex in operation, cannot realize batch detection, needs further calculation of results, is large in detection equipment volume and high in detection cost, and is difficult to popularize in daily life.
In order to achieve the purpose, the invention adopts the following technical scheme:
a tea amino acid content determination method based on colorimetric capsule image recognition comprises the following steps:
(1) preparing a color reaction reagent phosphate buffer solution and a ninhydrin solution, freezing and solidifying the color reaction reagent phosphate buffer solution and the ninhydrin solution into a plurality of unit blocks with fixed quantity respectively, wherein the solidified buffer solution unit block is called a capsule 1, the solidified ninhydrin solution unit block is called a capsule 2, and the color reaction reagent phosphate buffer solution and the ninhydrin solution are independently frozen and preserved for later use;
(2) adding a tea sample to be detected into boiling water for leaching to prepare a sample liquid to be detected;
(3) melting the capsule 1 and the capsule 2, adding the melted capsules into a colorimetric measuring cup to perform color reaction with a sample solution to be measured, and performing constant volume to obtain a sample color developing solution;
(4) photographing and collecting a color development image of a sample in the colorimetric measuring cup, and extracting an RGB value of the sample image by utilizing Python software to obtain sample image data;
(5) inputting the image data into a standard concentration model, and outputting the amino acid predicted concentration value of the sample to be detected in an operating mode;
the method for constructing the standard concentration model comprises the following steps:
a. preparing a series of standard solutions with concentration by using glutamic acid or theanine;
b. obtaining a standard solution image dataset according to the method of the steps (3) to (4);
c. dividing a standard liquid image data set into a training set and a testing set;
d. and respectively inputting the training set, the testing set and the corresponding concentrations into a learning model for training, and establishing the standard concentration model.
The tea can be various tea leaves on the market, such as green tea, black tea, matcha and the like.
According to the method, a mathematical model with the correlation between image information and concentration is constructed by acquiring the image information of color reaction liquid based on the color reaction between free amino acid in tea and ninhydrin. When the method is applied, the image data of the color reaction liquid of the tea sample to be detected is input into the mathematical model, and the predicted concentration value can be obtained.
In the step (1), the chromogenic reaction reagent is prepared into capsules with fixed amount per part, and the capsules are frozen and stored for later use, so that the step of preparing the reaction reagent each time is saved. During detection, one capsule 1 and one capsule 2 are consumed by one sample, and the color reaction can be generated by mixing and heating, so that the experimental steps are simplified.
Further, in step (1), phosphate buffer solution and color reagent 2% ninhydrin solution of pH8.0 are frozen at-20 deg.C into 0.5mL unit cell units of one volume, which are referred to as capsule 1 and capsule 2, respectively.
In the step (2), the preparation of the sample solution to be detected refers to the determination of the total amount of the GBT8314-2013 tea free amino acids.
Further, in the step (3), adding the melted buffer capsule 1 into a colorimetric measuring cup, adding 1mL of the sample solution to be detected, uniformly mixing, adding the melted developer capsule 2, heating in a boiling water bath for 15-20 minutes, cooling, and fixing the volume to obtain the sample detection color developing solution.
The colorimetric measuring cup is provided with a constant volume scale mark, and purified water is added after color reaction to achieve constant volume, so that the experimental consistency is ensured.
Further, in the step (4), the color development images of different samples are collected under the same environment and parameter conditions, so that the parameters of the collected images are ensured to be consistent. The invention is not limited to image acquisition devices and environments.
Further, in the step (4), blank correction was performed on the sample developed image using the standard solution developed image having an amino acid concentration of 0 mg/mL. All images were blank-corrected with a reference color image of 0mg/mL amino acid concentration. The purpose is to make the image color development clear and ensure the consistency of image acquisition.
Further, in step (4), RGB is converted into gray data using a gray formula.
Further, in the step (5), the predicted concentration output by the model is substituted into the following formula to be converted into the predicted content of the amino acid in the unit volume of the sample,
predicted concentration V/m/(1-water content)/1000X 100%
Wherein: predicted content of X-amino acid; v, determining the volume of the solution to be detected to be the volume with the unit of mL; m-the weight of the tea sample, the unit is g; the water content is generally 5% by default, and can be changed into a known water content value.
When a standard concentration model is established, glutamic acid or theanine is used as amino acid standard solution, and a series of standard solution with the concentration of 0-0.50mg/mL and the interval of 0.05mg/mL is prepared by referring to GBT 8314-2013. And performing color reaction, collecting color images, extracting RGB (red, green and blue) by the same method, and then establishing a model for a series of standard concentration image data to form a standard concentration RGB library.
Further, in step b, the number of image samples per density is more than or equal to 75.
In the model constructed by the invention, a data set is divided into a training set and a testing set according to a ratio of 9:1, image data is input into a nearest neighbor algorithm model, a random forest model and a Bayesian model, an ensemble learning model is established for the three models by using a Staking algorithm, and then a logistic regression classifier is used for secondary classification to obtain a standard concentration model. The accuracy of the obtained training set is 98.8, and the accuracy of the concentration classification result of the test set is 0.88. But the invention is not limited thereto.
The invention has the following beneficial effects:
(1) the invention prepares and stores the experimental reagent by a novel rapid reagent curing method through capsules, and saves the step of preparing the reaction reagent each time.
(2) The invention establishes a model with standard concentration corresponding to the image, predicts the amino acid content through image recognition, and when in application, the sample image information is introduced into the model, so that the amino acid content can be directly output, the step of re-making a standard curve for each detection is omitted, and the detection speed can be accelerated while the accuracy is ensured.
(3) The method of the invention does not limit the image acquisition equipment, reduces the instrument cost and does not need instrument training for operators.
(4) The method does not limit the image acquisition environment, breaks through the limit of the environment, and can realize the purpose of detection at any time and any place.
(5) The method can simultaneously and quickly carry out batch detection within the range of the running capacity of the equipment.
Drawings
FIG. 1 is a standard amino acid color image, in which 1 to 10 are concentrations of 0mg/mL,0.10mg/mL, 0.15mg/mL, 0.20mg/mL, 0.25mg/mL, 0.30mg/mL, 0.35mg/mL, 0.40mg/mL, 0.45mg/mL, and 0.50mg/mL, respectively.
Detailed Description
The present invention will be further illustrated with reference to the following examples, but the present invention is not limited thereto.
Example 1
1. The colorimetric capsule preparation technology comprises the following steps:
1.1 reference GBT8314-2013 tea free amino acid total amount determination buffer and color reagent preparation, specifically:
the procedure for preparing the phosphate buffer at pH8.0 was as follows: 23.9g of disodium hydrogen phosphate dodecahydrate (Na) were weighed2HPO4·12H2O), adding water to dissolve, transferring into a 1L volumetric flask, fixing the volume to the scale, and shaking up. Weighing potassium dihydrogen phosphate (KH) dried at 110 deg.C for 2 hr2PO4)9.08g of the powder is added with water to be dissolved and then transferred into a 1L volumetric flask, and the volume is fixed to the scale and shaken up. 95mL of the above disodium hydrogen phosphate solution and 5mL of potassium dihydrogen phosphate solution were mixed, and the pH of the mixture was 8.0.
Preparing a 2% ninhydrin solution: weighing 2g of ninhydrin (purity not less than 99%), adding 50mL of water and 80mg of stannous chloride (SnCl)2·2H2O) stirring evenly. Dissolving with small amount of water, standing in dark for a day and night, filtering, adding waterThe volume is 100 mL.
1.2 respectively placing the buffer solution and the color developing agent at the temperature of minus 20 ℃ to be frozen into 0.5mL of colorimetric capsule blocks with a volume unit, and respectively storing the colorimetric capsule blocks at the temperature of minus 20 ℃ for later use. Hereinafter referred to as a buffer unit as capsule 1. The developer unit is a capsule 2.
2. Establishing a standard amino acid concentration detection model, comprising the following steps:
2.1 taking L-glutamic acid (the purity is not lower than 99%) as amino acid standard solution, preparing a series of standard solutions with the concentration of 0-0.50mg/mL and the interval of 0.05 mg/mL. The preparation reference GBT8314 and 2013 is used for measuring the total amount of free amino acids in the tea.
2.2 taking the capsule 1 and the capsule 2, and melting at room temperature. The capsule 1 is put into a colorimetric measuring cup, 1mL of amino acid standard solution is added, and the mixture is shaken up. Adding the capsule 2, mixing, and shaking. Placing in boiled water for 15 minutes, cooling, and finally adding purified water to reach the constant volume to the scale mark.
2.3 shoot and gather the solution image after the development, specifically, the cell is steadily placed and is shot at the camera bellows middle part, and the light source is the LED shadowless lamp. The shooting parameters of the camera are M-gear, a shutter 1/200, an aperture f 5.6 and ISO 100, and the parameters of the collected images are ensured to be consistent.
The image processing comprises the following specific steps: and (3) introducing the images into PS software, carrying out white balance treatment by taking an amino acid standard solution with the concentration of 0mg/mL as a reference, and storing each concentration image of the sample after white balance.
A series of colorimetric images of amino acids at concentrations are shown in FIG. 1, and the concentrations from left to right are 0mg/mL,0.10mg/mL, 0.15mg/mL, 0.20mg/mL, 0.25mg/mL, 0.30mg/mL, 0.35mg/mL, 0.40mg/mL, 0.45mg/mL, and 0.50mg/mL, respectively.
And 2.4, importing the image into Python software, extracting RGB values of various concentrations and carrying out gray processing. And establishing a model for a series of standard concentration image data to form a standard concentration RGB library.
The method comprises the following specific steps: and importing the colorimetric image into Python software to extract RGB. RGB is converted into gradation data using a gradation formula of gradation R0.299 + G0.587 + B0.114. And performing histogram equalization processing on the image, and performing one-dimensional processing on the two-dimensional data.
100 image data are acquired for each concentration, and a standard concentration data set is divided into a training set: test set 9: 1. A nearest neighbor algorithm (KNN) model, a Random Forest (RF) model, a Bayes (Bayes) model and a logistic regression model are respectively established. And the random forest sets the decision tree n to be 200, and the other models are model default parameters. The obtained prediction classification results are shown in tables 1 to 4, and the accuracy rates are 0.80, 0.87, 0.68 and 0.68 respectively. And integrating the first three classifier models by using the starting, and performing secondary classification by using logistic regression. The results are shown in table 5, with a training set accuracy of 98.8 and a test set concentration classification result accuracy of 0.88.
TABLE 1 Standard concentration prediction Classification results based on nearest neighbor Algorithm (KNN) model
| Concentration of | Rate of accuracy | Recall rate | f1-score | Number of samples |
| 0 | 0.77 | 1.00 | 0.87 | 10 |
| 0.15 | 0.88 | 0.70 | 0.78 | 10 |
| 0.20 | 0.88 | 0.70 | 0.78 | 10 |
| 0.25 | 0.82 | 0.90 | 0.86 | 10 |
| 0.30 | 0.89 | 0.80 | 0.84 | 10 |
| 0.35 | 0.64 | 0.70 | 0.67 | 10 |
| 0.40 | 0.78 | 0.70 | 0.74 | 10 |
| 0.45 | 0.71 | 1.00 | 0.83 | 10 |
| 0.50 | 1.00 | 0.70 | 0.82 | 10 |
| Rate of accuracy | 0.80 | 90 |
TABLE 2 Standard concentration prediction Classification results based on Random Forest (RF) model
| Concentration of | Rate of accuracy | Recall rate | f1-score | Number of samples |
| 0 | 0.91 | 1.00 | 0.95 | 10 |
| 0.15 | 1.00 | 0.80 | 0.89 | 10 |
| 0.20 | 0.82 | 0.90 | 0.86 | 10 |
| 0.25 | 0.88 | 0.70 | 0.78 | 10 |
| 0.30 | 0.83 | 1.00 | 0.91 | 10 |
| 0.35 | 1.00 | 0.80 | 0.89 | 10 |
| 0.40 | 0.75 | 0.90 | 0.82 | 10 |
| 0.45 | 0.71 | 1.00 | 0.87 | 10 |
| 0.50 | 1.00 | 0.70 | 0.82 | 10 |
| Rate of accuracy | 0.87 | 90 |
TABLE 3 Standard concentration prediction Classification results based on Bayes model
TABLE 4 Standard concentration prediction Classification based on Logistic Regression (LR) model
| Concentration of | Rate of accuracy | Recall rate | f1-score | Number of samples |
| 0 | 0.62 | 1.00 | 0.77 | 10 |
| 0.15 | 1.00 | 0.30 | 0.46 | 10 |
| 0.20 | 0.77 | 1.00 | 0.87 | 10 |
| 0.25 | 1.00 | 0.60 | 0.75 | 10 |
| 0.30 | 0.64 | 0.90 | 0.75 | 10 |
| 0.35 | 0.44 | 0.40 | 0.42 | 10 |
| 0.40 | 0.50 | 0.40 | 0.44 | 10 |
| 0.45 | 0.62 | 0.80 | 0.70 | 10 |
| 0.50 | 0.88 | 0.70 | 0.78 | 10 |
| Rate of accuracy | 0.68 | 90 |
TABLE 5 Standard concentration prediction Classification results
Note: the precision ratio is as follows: correctly predicting as being in proportion to all predictions as being positive; the recall ratio is as follows: correctly predicting the proportion of the positive samples in the total positive samples; f 1-score: precision and recall are 2 times the harmonic mean.
From the prediction classification result, the prediction effect of each concentration is good, and the overall classification accuracy reaches 0.88.
Example 2
Black tea sample concentration prediction
6 black tea samples were selected: the tea is prepared from the following raw materials, namely inspection of special grade of Wangfu tea, inspection of first grade of Wangfu tea, inspection of special grade of Hosta tea (canning), inspection of special grade of Hosta tea (bagging), inspection of fine product of Hosta tea and inspection of Hosta tea (bagging).
The detection of the sample is carried out according to the following steps:
1. and (4) preparing a sample solution to be tested by referring to the measurement of the total amount of the GBT8314-2013 tea free amino acids.
Specifically, 3g (to 0.001g) of ground tea sample is weighed into a 500mL conical flask (or prepared in proportion), 450mL of boiling distilled water is added, the mixture is immediately transferred into a water bath, leaching is carried out for 45min (shaking is carried out every 10 min), pressure reduction filtration is carried out immediately after leaching is finished while the mixture is hot, and residues are washed for 2-3 times by a small amount of hot distilled water. Transferring the filtrate into a 500mL volumetric flask, cooling, adding water to a constant volume to a scale, and shaking up.
2. Melting the capsule 1 and the capsule 2 at room temperature. And putting the capsule 1 into a colorimetric measuring cup, adding 1mL of a solution sample to be measured, and shaking up. Adding the capsule 2, mixing, and shaking. Placing in boiled water for 15 minutes, cooling, and finally adding purified water to reach the constant volume to the scale mark.
3. Photographing and collecting a color development image of a sample solution, wherein the image processing comprises the following specific steps: the images were introduced into PS software and subjected to white balance treatment based on an amino acid standard solution having a concentration of 0 mg/mL. And storing each sample image of the white-balanced image layer of the sample. And importing the sample image into Python software to extract RGB values and perform gray processing to obtain sample image data. The procedure was as in example 1.
4. The image data was introduced into the standard concentration model established in example 1 to obtain the concentration of the sample. And then the predicted content of the amino acid in the unit volume sample is obtained through conversion according to a formula. The predicted amino acid content X (%) was calculated using the following formula:
predicted concentration V/m/(1-water content)/1000 100
Wherein: v, determining the volume of the solution to be detected to be the volume with the unit of mL;
m-the weight of the tea sample, the unit is g;
the water content is 5% by default, and can also be calculated according to the actual known quantity.
Inputting the weighed mass and the water content of the sample, and outputting the amino acid content value of the sample in unit volume. Here, three replicates of each sample were averaged.
Predicted value of amino acid content of black tea is as follows:
TABLE 6 comparison of content of measured and predicted values of black tea samples
| Sample (I) | Measured value (%) | Predicted value (%) |
| Special grade Wangfu tea | 4.38 | 4.59 |
| Wangfu tea first grade | 4.55 | 4.86 |
| Special tea Wang Hai (canned) | 3.65 | 3.81 |
| Special grade Wanghai tea (bag package) | 3.35 | 3.63 |
| Wanghai tea extract | 3.93 | 4.01 |
| Hope sea tea (bag package) | 3.74 | 3.80 |
Note: the measured value is determined by referring to the total amount of free amino acids in GBT8314-2013 tea. The same applies below.
Predicting the amino acid content of 6 black tea samples, R2And the predicted effect is better and the stability is better when the RMSEP is 0.091 and 0.757.
Example 3
Green tea sample concentration prediction
6 green tea samples were selected: the special grade of the Huangshan Maofeng is first grade, the special grade of the Huangshan Maofeng is second grade, the special grade of the Huangshan Maofeng is third grade, the Huangshan Maofeng is first grade, the Huangshan Maofeng is second grade and the Huangshan Maofeng is third grade.
The detection of the sample is carried out according to the following steps:
1. and (4) preparing a sample solution to be tested by referring to the measurement of the total amount of the GBT8314-2013 tea free amino acids.
2. Melting the capsule 1 and the capsule 2 at room temperature. And putting the capsule 1 into a colorimetric measuring cup, adding 1mL of a solution sample to be measured, and shaking up. Adding the capsule 2, mixing, and shaking. Placing in boiled water for 15 minutes, cooling, and finally adding purified water to reach the constant volume to the scale mark.
3. Photographing and collecting a color development image of a sample solution, wherein the image processing comprises the following specific steps: the images were introduced into PS software and subjected to white balance treatment based on an amino acid standard solution having a concentration of 0 mg/mL. And storing each sample image of the white-balanced image layer of the sample. And importing the sample image into Python software to extract RGB values and perform gray processing to obtain sample image data.
4. The image data was introduced into the standard concentration model established in example 1 to obtain the concentration of the sample. And then the predicted content of the amino acid in the solution per unit volume is obtained through conversion according to a formula. The predicted amino acid content X (%) was calculated using the following formula:
predicted concentration V/m/(1-water content)/1000 100
Wherein: v, determining the volume of the solution to be detected to be the volume with the unit of mL;
m-the weight of the tea sample, the unit is g;
the water content is 5% by default, and can also be calculated according to the actual known quantity.
Inputting the weighing mass and the water content of the green tea sample, and outputting the amino acid content value of the sample in unit volume. Here, three replicates of each sample were averaged.
The predicted value of the amino acid content of the green tea is as follows:
TABLE 7 comparison of measured and predicted values for green tea samples
| Sample (I) | Measured value (%) | Predicted value (%) |
| Sheyu rhubarb mountain Maofeng super grade | 4.35 | 4.42 |
| Sheyu rhubarb mountain Maofeng special grade two | 4.85 | 5.03 |
| Sheyu rhubarb mountain Maofeng super grade III | 4.84 | 5.05 |
| Sheyu rhubarb mountain Maofeng | 5.26 | 5.36 |
| Second grade of Sheyu Dahuangshan Maofeng | 4.81 | 4.80 |
| Three-stage of Sheyu Dahuangshan Maofeng | 4.81 | 5.19 |
Amino acid content prediction, R, was performed on 6 green tea samples2The RMSEP is 0.792, and the RMSEP is 0.088, so the prediction effect is better, and the stability is better.
Example 4
Predicting the concentration of the matcha sample, comprising the following steps:
selecting 6 matcha samples: xiangyu matcha jianwu group, Xiangyu matcha haibei, Xiangyu matcha chunliu I, matcha B B type longjing 43, matcha type maolang powder and matcha longjing 43 #.
The detection of the sample is carried out according to the following steps:
1. and (4) preparing a sample solution to be tested by referring to the measurement of the total amount of the GBT8314-2013 tea free amino acids.
2. Melting the capsule 1 and the capsule 2 at room temperature. And putting the capsule 1 into a colorimetric measuring cup, adding 1mL of a solution sample to be measured, and shaking up. Adding the capsule 2, mixing, and shaking. Placing in boiled water for 15 minutes, cooling, and finally adding purified water to reach the constant volume to the scale mark.
3. Photographing and collecting a color development image of a sample solution, wherein the image processing comprises the following specific steps: the images were introduced into PS software and subjected to white balance treatment based on an amino acid standard solution having a concentration of 0 mg/mL. And storing each sample image of the white-balanced image layer of the sample. And importing the sample image into Python software to extract RGB values and perform gray processing to obtain sample image data.
4. The image data was introduced into the standard concentration model established in example 1 to obtain the concentration of the sample. And then the predicted content of the amino acid in the solution per unit volume is obtained through conversion according to a formula. The predicted amino acid content X (%) was calculated using the following formula:
predicted concentration V/m/(1-water content)/1000 100
Wherein: v, determining the volume of the solution to be detected to be the volume with the unit of mL;
m-the weight of the tea sample, the unit is g;
the water content is 5% by default, and can also be calculated according to the actual known quantity.
Inputting the weight and the water content of the matcha sample, and outputting the amino acid content value of the sample in unit volume. Here, three replicates of each sample were averaged.
The predicted value of the amino acid content of matcha is shown in the following table:
TABLE 8 comparison of real value and predicted value of matcha sample
| Sample (I) | Measured value (%) | Predicted value (%) |
| Rural rain matcha dove-hole group | 3.68 | 3.42 |
| Township Iyu Matcha | 4.49 | 4.61 |
| Village rain matcha spring rain No. I | 5.26 | 6.03 |
| Matcha B B-like Longjing 43 | 3.78 | 3.65 |
| Matcha A-like luo-green powder | 5.28 | 5.66 |
| Matcha Longjing 43# | 4.36 | 4.86 |
Predicting the amino acid content of 6 matcha samples, R2When the RMSEP is equal to 0.539 and the RMSEP is equal to 0.191, the predicted effect is general, and the stability is better.
Claims (9)
1. A method for measuring the amino acid content of tea leaves based on colorimetric capsule image recognition is characterized by comprising the following steps:
(1) preparing a color reaction reagent phosphate buffer solution and a ninhydrin solution, freezing and solidifying the color reaction reagent phosphate buffer solution and the ninhydrin solution into a plurality of unit blocks with fixed quantity respectively, wherein the solidified buffer solution unit block is called a capsule 1, the solidified ninhydrin solution unit block is called a capsule 2, and the color reaction reagent phosphate buffer solution and the ninhydrin solution are independently frozen and preserved for later use;
(2) adding a tea sample to be detected into boiling water for leaching to prepare a sample liquid to be detected;
(3) melting the capsule 1 and the capsule 2, adding the melted capsules into a colorimetric measuring cup to perform color reaction with a sample solution to be measured, and performing constant volume to obtain a sample color developing solution;
(4) photographing and collecting a color development image of a sample in the colorimetric measuring cup, and extracting an RGB value of the sample image by utilizing Python software to obtain sample image data;
(5) inputting the image data into a standard concentration model, and outputting the amino acid predicted concentration value of the sample to be detected in an operating mode;
the method for constructing the standard concentration model comprises the following steps:
a. preparing a series of standard solutions with concentration by using glutamic acid or theanine;
b. obtaining a standard solution image dataset according to the method of the steps (3) to (4);
c. dividing a standard liquid image data set into a training set and a testing set;
d. and respectively inputting the training set, the testing set and the corresponding concentrations into a learning model for training, and establishing the standard concentration model.
2. The method for determining the amino acid content in tea leaves based on colorimetric capsule image recognition according to claim 1, wherein in the step (1), the phosphate buffer solution with pH of 8.0 and the 2% ninhydrin solution as a color developing agent are respectively frozen into 0.5mL of capsules with one volume unit.
3. The method for measuring the amino acid content in the tea leaves based on the colorimetric capsule image recognition according to claim 2, wherein in the step (3), the melted capsule 1 is added into a colorimetric measuring cup, 1mL of a sample solution to be measured is added, the melted capsule 2 is added after being mixed uniformly, the mixture is heated in a boiling water bath for 15 minutes, and the volume is determined after the mixture is cooled, so that a sample color developing solution is obtained.
4. The method for determining amino acid content in tea leaves based on colorimetric capsule image recognition according to claim 1, wherein in the step (4), color development images of different samples are collected under the same conditions.
5. The method for determining the amino acid content in tea leaves based on colorimetric capsule image recognition according to claim 1, wherein in the step (4), the sample developed image is blank-corrected by a standard solution developed image with an amino acid concentration of 0 mg/mL.
6. The method for measuring amino acid content in tea leaves based on colorimetric capsule image recognition according to claim 1, wherein in the step (4), RGB is converted into gray data by using a gray formula.
7. The method for determining the amino acid content in the tea leaves based on the colorimetric capsule image recognition according to claim 1, wherein in the step b, the number of image samples for each standard concentration for modeling is not less than 75.
8. The method for measuring the amino acid content in the tea leaves based on the colorimetric capsule image recognition as claimed in claim 1, wherein in the step d, the training set and the test set and the corresponding concentrations are respectively input into a nearest neighbor algorithm model, a random forest model and a Bayesian model, an ensemble learning model is built for the three models by using a labeling algorithm, and then a logistic regression classifier is used for secondary classification to obtain a standard concentration model.
9. The method for determining the amino acid content of tea leaves based on colorimetric capsule image recognition according to claim 1, wherein the tea leaves are green tea, black tea, matcha.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011241593.2A CN112461820B (en) | 2020-11-09 | 2020-11-09 | Tea amino acid content determination method based on colorimetric capsule image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011241593.2A CN112461820B (en) | 2020-11-09 | 2020-11-09 | Tea amino acid content determination method based on colorimetric capsule image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112461820A true CN112461820A (en) | 2021-03-09 |
| CN112461820B CN112461820B (en) | 2022-02-11 |
Family
ID=74826321
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011241593.2A Active CN112461820B (en) | 2020-11-09 | 2020-11-09 | Tea amino acid content determination method based on colorimetric capsule image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112461820B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150055134A1 (en) * | 2012-02-03 | 2015-02-26 | University Of Cincinnati | Method and system for analyzing a colorimetric assay |
| CN105675507A (en) * | 2016-01-18 | 2016-06-15 | 浙江大学 | Mobile phone color analysis based cysteine detection method |
| CN106353311A (en) * | 2016-08-30 | 2017-01-25 | 开平市中青环保技术服务有限公司 | Intelligent detection method and device for realizing material qualitative for using visual sensor detection based on colorimetric reaction |
| US20170026620A1 (en) * | 2015-07-24 | 2017-01-26 | Venkatasubramaniam Kalambur | Apparatus and a method to determine the concentration of an analyte |
| CN107807098A (en) * | 2017-10-25 | 2018-03-16 | 江苏海德环境科技有限公司 | A kind of heavy metal colorimetric identification device and method based on image procossing |
| CN109323999A (en) * | 2018-10-24 | 2019-02-12 | 大连理工大学 | A kind of spectrophotometric detecting method based on image values analysis |
| CN109374607A (en) * | 2018-10-18 | 2019-02-22 | 湖南农业大学 | The measuring method of free amino acid total amount in a kind of tea beverage |
| CN110632067A (en) * | 2019-10-29 | 2019-12-31 | 饶宾期 | Machine vision-based water body dissolved oxygen detection method |
| CN110672784A (en) * | 2019-10-29 | 2020-01-10 | 饶宾期 | Water body dissolved oxygen detection device based on machine vision |
| CN110763674A (en) * | 2019-12-03 | 2020-02-07 | 淮北师范大学 | Method for rapidly detecting content of vitamin C in vegetables and fruits |
| CN111239119A (en) * | 2020-02-25 | 2020-06-05 | 西安交通大学 | System and method for colorimetric detection of mercury ions based on App |
| CN111272746A (en) * | 2020-03-09 | 2020-06-12 | 中国水产科学研究院东海水产研究所 | Method for determining total amount of free amino acids in aquatic product by colorimetric method |
| CN111426761A (en) * | 2020-03-19 | 2020-07-17 | 东莞市中鼎检测技术有限公司 | Method for identifying Tieguanyin tea leaves in different years |
-
2020
- 2020-11-09 CN CN202011241593.2A patent/CN112461820B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150055134A1 (en) * | 2012-02-03 | 2015-02-26 | University Of Cincinnati | Method and system for analyzing a colorimetric assay |
| US20170026620A1 (en) * | 2015-07-24 | 2017-01-26 | Venkatasubramaniam Kalambur | Apparatus and a method to determine the concentration of an analyte |
| CN105675507A (en) * | 2016-01-18 | 2016-06-15 | 浙江大学 | Mobile phone color analysis based cysteine detection method |
| CN106353311A (en) * | 2016-08-30 | 2017-01-25 | 开平市中青环保技术服务有限公司 | Intelligent detection method and device for realizing material qualitative for using visual sensor detection based on colorimetric reaction |
| CN107807098A (en) * | 2017-10-25 | 2018-03-16 | 江苏海德环境科技有限公司 | A kind of heavy metal colorimetric identification device and method based on image procossing |
| CN109374607A (en) * | 2018-10-18 | 2019-02-22 | 湖南农业大学 | The measuring method of free amino acid total amount in a kind of tea beverage |
| CN109323999A (en) * | 2018-10-24 | 2019-02-12 | 大连理工大学 | A kind of spectrophotometric detecting method based on image values analysis |
| CN110632067A (en) * | 2019-10-29 | 2019-12-31 | 饶宾期 | Machine vision-based water body dissolved oxygen detection method |
| CN110672784A (en) * | 2019-10-29 | 2020-01-10 | 饶宾期 | Water body dissolved oxygen detection device based on machine vision |
| CN110763674A (en) * | 2019-12-03 | 2020-02-07 | 淮北师范大学 | Method for rapidly detecting content of vitamin C in vegetables and fruits |
| CN111239119A (en) * | 2020-02-25 | 2020-06-05 | 西安交通大学 | System and method for colorimetric detection of mercury ions based on App |
| CN111272746A (en) * | 2020-03-09 | 2020-06-12 | 中国水产科学研究院东海水产研究所 | Method for determining total amount of free amino acids in aquatic product by colorimetric method |
| CN111426761A (en) * | 2020-03-19 | 2020-07-17 | 东莞市中鼎检测技术有限公司 | Method for identifying Tieguanyin tea leaves in different years |
Non-Patent Citations (5)
| Title |
|---|
| DANYI HUANG ET AL.: "Rapid Identification of Different Grades of Huangshan Maofeng Tea Using Ultraviolet Spectrum and Color Difference", 《MOLECULES》 * |
| SILVIJA ŠAFRANKO ET AL.: "Designing ColorX, Image Processing Software for Colorimetric Determination of Concentration, To Facilitate Students’ Investigation of Analytical Chemistry Concepts Using Digital Imaging Technology", 《JOURNAL OF CHEMICAL EDUCATION》 * |
| 中华人民共和国国家质量监督检验检疫总局 等: "《中华人民共和国国家标准GB/T8314-2013 茶 游离氨基酸总量的测定》", 31 December 2013, 中国标准出版社 * |
| 刘锐 等: "图像比色法测定有色溶液", 《理化检验-化学分册》 * |
| 沈继忱 等: "基于图像比色法的有色溶液浓度检测方法的研究", 《光学仪器》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112461820B (en) | 2022-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Podio et al. | Elemental and isotopic fingerprint of Argentinean wheat. Matching soil, water, and crop composition to differentiate provenance | |
| CN108680572B (en) | Method for visually and rapidly detecting nitrite by using color/fluorescence dual signals and application | |
| Cammack et al. | Fluorescent dissolved organic matter in lakes: relationships with heterotrophic metabolism | |
| CN105181912B (en) | A kind of Noninvasive Measuring Method of Freshness in rice storage | |
| CN105699472B (en) | Flat tea place of production method of discrimination based on stable isotope proportional difference | |
| CN109323999B (en) | Spectrophotometric detection method based on image numerical analysis | |
| CN101806719A (en) | Method for measuring content of available phosphorus in soil | |
| CN113008817A (en) | Method for rapidly identifying authenticity and quality of bitter apricot kernels based on hyperspectral imaging technology | |
| CN103063663A (en) | Testing method for total phosphorus content in soil based on image analysis | |
| CN112461820B (en) | Tea amino acid content determination method based on colorimetric capsule image recognition | |
| CN110596080A (en) | Mineral element-based golden pomfret origin identification method | |
| Choi et al. | Semi-quantitative analysis for selecting Fe-and Zn-dense genotypes of staple food crops | |
| CN111024665A (en) | Method for measuring content of organic selenium in selenium-enriched rice or product prepared from selenium-enriched rice | |
| US7371565B2 (en) | Apparatus and method for measuring intracellular reactions | |
| CN112200446A (en) | Method for comprehensively evaluating pork quality | |
| CN106442500A (en) | Raw milk bacterium concentration detection method | |
| CN111239070A (en) | Method for measuring uniform mixing degree of raw materials of pellet mixer | |
| Xiong et al. | Authenticating the geographical origin of the Chinese yam (Tiegun) with stable isotopes and multiple elements | |
| CN112924448B (en) | Determination method for detecting true protein content by salting-out method and application thereof | |
| CN112326647A (en) | Phosphorus content detection reagent based on molybdenum trioxide and preparation method and detection method thereof | |
| CN113030080A (en) | Method for evaluating freshness of rice in storage process | |
| CN109856063A (en) | The detection method and system of synthetic dyestuff concentration in soda | |
| CN110376175A (en) | A kind of synthesis fluorescent carbon point is used for selective enumeration method Hg2+Method | |
| CN109521011A (en) | The testing cassete of available phosphorus content and the method using available phosphorus in the testing cassete testing soil in a kind of portable quick detection soil | |
| Su et al. | Assessing geographical origin of Diqing wines based on their elemental and isotopic profiles |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |