Synchronous rapid detection method for contents of tea polyphenol and caffeine in jasmine tea

Abstract

The invention discloses a method for synchronously and rapidly detecting the contents of tea polyphenol and caffeine in jasmine tea, which comprises the following steps of: measuring the brightness (L), the green-red degree (a), the blue-yellow degree (b), the total color difference (E) and other characteristic values of different dry tea samples by using a spectrophotometer; then, brewing and extracting with boiling water to obtain tea soup, putting the tea soup into a self-made image acquisition camera obscura, acquiring a tea soup image, and extracting a tea soup image characteristic value by using MATLAB (Matrix Laboratory); and inputting the obtained color difference of the dry tea and the characteristic value of the tea soup image into the constructed BP neural network prediction model, and respectively calculating the contents of the tea polyphenol and the caffeine in the sample. The method disclosed by the invention can be used for simultaneously detecting two indexes of tea polyphenol and caffeine, is simple and convenient to operate, does not need to use chemical reagents, is green and environment-friendly, is only 7.76% of the time consumed by a detection method of a traditional national standard method, and greatly improves the detection efficiency.

Description

Technical Field

[0001] The present invention relates to the field of chemical analysis technology, and aims to provide a method for simply, quickly and accurately detecting the content of tea polyphenols and caffeine in jasmine tea. This method is not only applicable to jasmine tea, but also to other types of tea, and provides an effective means for controlling and evaluating the quality of tea. Background Art

[0002] At present, the main indicators for the quality of jasmine tea are sensory indicators, physical and chemical indicators, and aroma indicators. Among them, tea polyphenols in the physical and chemical indicators are one of the main active substances in jasmine tea, and have multiple biological activities such as lowering blood sugar and blood lipids. The current tea polyphenol detection method is the national standard GB / T8313-2018 Folin phenol method, but this method is a chemical analysis method, which is cumbersome, time-consuming, uses more organic reagents, and is not environmentally friendly. Caffeine in the physical and chemical indicators is another main active substance in jasmine tea, which has a significant anti-inflammatory effect. At present, jasmine tea caffeine is mainly determined by the national standard GB / T8312-2013 method, but this method is also a chemical analysis method and has similar shortcomings as mentioned above.

[0003] The harvest period of jasmine is only 3 months, and the storage period is long. During the storage period, it is easily affected by the external environment, which leads to a serious decline in product quality. Therefore, it is necessary to develop a simple, rapid and sensitive method for rapid detection of tea polyphenols and caffeine content in jasmine tea, which is suitable for on-site rapid detection, environmentally friendly, and can meet the needs of the tea products industry and consumers for rapid detection of typical physical and chemical indicators of jasmine tea. Summary of the invention

[0004] The present invention provides a method for synchronously and rapidly detecting the content of tea polyphenols and caffeine in jasmine tea. The method comprises the following steps: firstly, the color difference characteristic value of a dry tea sample is measured by a colorimeter. Then, the tea soup is brewed and extracted with boiling water to obtain the tea soup, which is then put into a self-made "image acquisition dark box" to collect the tea soup image, and the characteristic value of the tea soup image is extracted by MATLAB. The obtained dry tea color difference and tea soup image characteristic value are input into a constructed BP neural network prediction model to calculate the content of tea polyphenols and caffeine in the sample respectively.

[0005] The method for synchronously and rapidly detecting the content of tea polyphenols and caffeine in jasmine tea of ​​the present invention comprises the following steps:

[0006] (1) Evenly spread a certain amount of dried jasmine tea samples on a white cardboard, pay attention to keep the samples even, avoid direct light exposure to the samples, use a colorimeter to measure at least one of the sample's L brightness, a green-red intensity, b blue-yellow intensity, and ΔE total color difference, each sample is measured three times in parallel, and the average value is taken;

[0007] (2) Accurately weigh 2-10 g of jasmine tea sample, brew it in boiling water at a ratio of 1:50-80 g / g (i.e., add 50-80 g of boiling water per gram of tea sample), let it stand for 3-10 minutes, put the prepared tea soup in a white porcelain bowl, cool it to about 20-50° C., put it in an "image acquisition dark box" to take pictures, and use MATLAB software to extract relevant feature values ​​of the tea soup image, wherein the relevant feature values ​​are one or more of H hue, I brightness, and S saturation color.

[0008] (3) The obtained relevant dry tea color difference and tea soup image feature values ​​are respectively input into the constructed BP neural network prediction model of jasmine tea polyphenols and caffeine to calculate the content of tea polyphenols and caffeine in the sample.

[0009] Preferably, in step (1), the green-redness and ΔE total color difference characteristic value of sample a are measured using a colorimeter;

[0010] Preferably, in step (2), 3 g of tea sample is weighed, brewed in boiling water at a ratio of 1:50 g / ml, and allowed to stand for 5 min. The sample is placed in a self-made "closed image acquisition box" containing an LED downlight (power 7w, model ESTTLT1391), and photographed and collected by a smartphone with a pixel of 48 million or more;

[0011] Preferably, the tea soup image-related feature values ​​are calculated after image denoising and image enhancement are performed on the photo taken in step (2); further preferably, median filtering is used for image denoising, and histogram equalization is used for image enhancement.

[0012] Preferably, the relevant characteristic value described in step (2) is I brightness.

[0013] Preferably, the color difference of the relevant dry tea and the tea soup image feature values ​​obtained in step (3) are respectively input into the constructed BP neural network prediction model of jasmine tea polyphenols and caffeine, and the BP neural network prediction model optimized by genetic algorithm is preferred, which has higher accuracy; the optimal number of hidden layer nodes of the neural network prediction model of tea polyphenols and caffeine is preferably 7 or 4. The training set mean square error (MSE) and the predicted error value of the established quantitative prediction model of jasmine tea caffeine and tea polyphenols are both small, and the mean square error is 0.0000114 and 0.00014665 respectively.

[0014] Further preferably, the BP prediction model for tea polyphenol content in step (3) is determined as: tea polyphenol BP neural network model: the input layer is 2, which are the original tea color difference characteristics a green-red intensity and △E total color difference characteristic value, the hidden layer is 1, the number of nodes is 7, and the output layer is 1 (i.e. tea polyphenol).

[0015] Further preferably, the caffeine BP neural network model has: an input layer of 3, which are the original tea color difference characteristics a green-red intensity and △E total color difference characteristic value and tea soup color I brightness, a hidden layer of 1, a node number of 4, and an output layer of 1 (ie caffeine).

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] (1) The method of the present invention does not require the use of any chemical reagents, is environmentally friendly, and is simple to operate;

[0018] (2) The method of the present invention can process a large number of samples at a time, and the relevant data collected by the samples can be used for the detection of tea polyphenols and caffeine at the same time. Excluding the cooling time of about 30 minutes after brewing the tea soup, the detection of the two physical and chemical indicators of tea polyphenols and caffeine of multiple samples can be completed synchronously within 1-2 minutes. This method greatly improves the detection efficiency and can be used for rapid on-site detection;

[0019] (3) The average absolute errors of the detection of the two substances by the method of the present invention are controlled within 5%. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is the test result of dry matter, free amino acids, tea polyphenols and caffeine content of different jasmine tea samples;

[0021] Figure 2 This is a correlation analysis chart between the color difference characteristics of jasmine tea dry tea and the tea polyphenols and caffeine content of the tea leaves;

[0022] Figure 3 This is a correlation analysis diagram between the image features of jasmine tea soup and the tea polyphenols and caffeine content of the tea leaves;

[0023] Figure 4 Comparison chart of the error (A) and index (B) between the BP neural network prediction value and the true value before and after optimization of jasmine tea polyphenols.

[0024] Figure 5 Comparison chart of the error (A) and index (B) between the BP neural network prediction value and the true value before and after the optimization of caffeine in jasmine tea. DETAILED DESCRIPTION

[0025] The terms used in the present invention, unless otherwise specified, generally have the meanings commonly understood by those of ordinary skill in the art. The present invention will be further described in detail below in conjunction with specific embodiments and with reference to data. It should be understood that these embodiments are only intended to illustrate the present invention and are not intended to limit the scope of the present invention in any way.

[0026] Experimental Materials

[0027] ① Jasmine tea sample A, named Chunlun Jasmine Red Gold (Fuzhou Jasmine Tea), the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0028] ② Jasmine tea sample B, named Longhao Jasmine (Jasmine tea), the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0029] ③ Jasmine tea sample C, named Zhang Yiyuan Jasmine Tea (Longhao), the tea base is Fuding white tea, produced in Hengxian County, Guangxi.

[0030] ④ Jasmine tea sample D, named Chunlun Jasmine Dabaihao (Fuzhou Jasmine Tea), the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0031] ⑤ Jasmine tea sample E, named Chunlun Jasmine Silver Snail (Jasmine Tea), the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0032] ⑥Jasmine tea sample F, named Zhang Yiyuan Jasmine Tea (Xiangxue), the tea base is roasted green tea, produced in Tongzhou District, Beijing.

[0033] ⑦ Jasmine tea sample G, named Chunlun Jasmine Yinhao (Jasmine Tea), the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0034] ⑧Jasmine tea sample H, named Chunlun Jasmine Yinhao (Jasmine tea), the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0035] ⑨Jasmine tea sample I, named Chunlun Fuzhou Jasmine Tea, the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0036] ⑩Jasmine tea sample J, named Chunlun Jasmine Dabaihao (Jasmine Tea), the tea base is roasted green tea, produced in Fuzhou, Fujian.

[0037] Jasmine tea sample K, named Chunlun Jasmine Tea (Longjing), is made of roasted green tea produced in Fuzhou, Fujian.

[0038] Jasmine tea sample L, named Fumingyuan Jasmine Silver Needle, is made from roasted green tea produced in Fuzhou, Fujian.

[0039] Jasmine tea sample M, named Monkey King Jasmine Tea (Chun Hao), is made of roasted green tea produced in Changsha, Hunan.

[0040] Jasmine tea sample N, named Zhang Yiyuan Jasmine Tea (Maojian), the tea base is roasted green tea, produced in Tongzhou, Beijing.

[0041] Jasmine tea sample O, named Monkey King Special Jasmine Tea No. 1, is made of roasted green tea produced in Changsha, Hunan.

[0042] Jasmine tea sample P is named Wu Yutai Special Jasmine Tea. The tea base is roasted green tea and is produced in Changping District, Beijing.

[0043] Testing methods for physical and chemical indicators of jasmine tea polyphenols, caffeine, etc.

[0044] (1) Determination of dry matter content. According to the national standard GB / T8304-2013, before conducting physical and chemical analysis on tea samples, the tea samples need to be crushed, sieved through an 80-mesh sieve, and stored at 4°C.

[0045] (2) Determination of free amino acids. Refer to the national standard GB / T8314-2013 for the determination of the total amount of free amino acids, i.e., the ninhydrin colorimetric method. Each sample was measured three times in parallel, and the average value and standard deviation were calculated.

[0046] (3) Determination of tea polyphenol content. Referring to the method for determining tea polyphenol content in the national standard GB / T8313-2018, tea polyphenols in the crushed tea sample were extracted using 70% methanol solution in a 70°C water bath. Gallic acid was used as the standard sample to draw a standard curve. The tea polyphenol content in the sample was quantitatively determined based on the absorbance of the sample at 765 nm. Each sample was measured three times in parallel, and the average value and standard deviation were calculated.

[0047] (4) Determination of caffeine content. The caffeine content in tea was determined by ultraviolet spectrophotometry according to the national standard GB / T8312-2013, and a caffeine standard curve was drawn. The caffeine content of the sample was quantitatively determined based on the absorbance of the sample at 274 nm. Each sample was measured three times in parallel, and the average value and standard deviation were calculated.

[0048] Jasmine tea color difference and tea soup image acquisition

[0049] (1) Color difference collection: First, a certain amount of tea sample was evenly spread on a white cardboard, and attention was paid to keeping the sample uniform and avoiding direct light exposure to the sample. The sample was measured using a spectrophotometer (NS820, Shenzhen Sanenshi Technology Co., Ltd.), and the L, a, b and △E of the sample were recorded. Each sample was measured three times in parallel and the average value was taken. Among them, L represents brightness (brightness), a represents the degree of green (red) when it is positive (negative) value; b represents the degree of blue (yellow) when it is negative (positive) value, b / a is the hue index, and △E represents the total color difference.

[0050] (2) Tea soup image acquisition: The evaluation method currently used in the national famous tea evaluation in my country is adopted, that is, a standard refined green tea evaluation cup and bowl is used, 3g of tea sample is accurately weighed, and brewed with boiling water at a ratio of 1:50. After standing for 5 minutes, the prepared tea soup is placed in a white porcelain evaluation bowl specially used for tea evaluation, cooled to about 30℃ (to prevent water vapor from interfering with image acquisition), and placed in a self-made "closed image acquisition box" containing an LED downlight (power 7W, model ESTTLT1391), and photos are collected using a smartphone (48 million pixels or above).

[0051] Tea soup image preprocessing

[0052] (1) Image denoising

[0053] The experiment uses the median filter method, which is a nonlinear signal processing method. Under certain conditions, it can not only avoid the blurring of image details caused by current filters such as minimum mean square filtering and mean filtering, but also has the best effect on filtering out pulse interference and image scanning noise. By removing isolated noise points, this method can also make the image less blurred. Image denoising is achieved based on MATLAB programming.

[0054] (2) Image enhancement

[0055] This experiment uses histogram equalization to enhance the image. This method transforms the histogram distribution of a given image into a uniform histogram distribution so that the probability density of the output pixel grayscale is evenly distributed. This increases the dynamic range of the image grayscale value, thereby achieving the effect of enhancing the overall contrast of the image and increasing the recognition accuracy of image features.

[0056] (3) Image feature value extraction

[0057] Tea soup image color feature extraction: When evaluating tea, the evaluation of tea soup includes the evaluation of tea soup color (i.e., soup color). Soup color is one of the important evaluation factors for tea sensory evaluation. Therefore, the color information of tea soup becomes the main characteristic of tea soup quality. In this study, the H hue, I brightness, and S saturation color system are still used to express the color characteristics of tea soup.

[0058] Evaluation model construction and verification

[0059] (1) Evaluation model construction

[0060] In this experiment, BP neural network was used to create a model for determining the content of tea polyphenols and caffeine in jasmine tea, and the BP neural network was optimized based on genetic algorithm (GN). BP neural network is a multi-layer feedforward neural network based on error back propagation. BP neural network includes input layer, hidden layer and output layer. In this experiment, the number of input layer units is determined according to the dimension of input information, that is, the relevant eigenvalues ​​significantly related to the physical and chemical indicators of tea are screened by using the principle of Pearson correlation analysis. If there are two eigenvalues ​​that are significantly correlated, the number of units is set to 2, and so on; the number of output layer units is 1, which is one of the physical and chemical indicators of tea, caffeine or tea polyphenols. The hidden layer is set to 1. The number of nodes in the hidden layer will affect the accuracy of model recognition, so this experiment selects the number of nodes with the smallest root mean square error of the training set. The S-type function is used as the action function of the model; the learning factor and momentum factor in the training process are both set to 0.1. This experiment uses the neural network toolbox in MATLAB to train the network. However, BP neural network has problems such as taking local extreme values ​​(optimal solutions) and low training efficiency. Genetic algorithms can make up for the problem of BP neural network finding the optimal solution. Its principle is similar to biological genetic evolution, that is, "survival of the fittest", and screen out the optimal solution.

[0061] (2) Model verification

[0062] ①Reliability test of the model

[0063] This experiment uses mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), and mean absolute percentage error (MAPE) as indicators to evaluate the prediction effect of the model. The smaller the MAE, MSE, RMSE, and MAPE, the more reliable the established model is.

[0064] ②Model accuracy experiment

[0065] Accuracy is one of the basic indicators for evaluating model performance. The error between the predicted value and the measured value can be used to reflect the accuracy. The smaller the error, the higher the accuracy.

[0066] Example 1: Determination of dry matter, free amino acids, tea polyphenols and caffeine content in different jasmine teas

[0067] The results of dry matter, free amino acids, tea polyphenols and caffeine content determination of different jasmine samples are as follows Figure 1 shown.

[0068] (1) Dry matter content

[0069] Depend on Figure 1 (A) It can be seen that the dry matter content of the 16 types of jasmine tea used in this experiment is within the range of 90% to 94%. According to the results of variance analysis, the dry matter content of the 16 types of jasmine tea has certain differences and a high degree of differentiation.

[0070] (2) Free amino acid content

[0071] Amino acid standard curve regression equation: y = 0.03525x-0.1118, correlation coefficient R 2 =0.99921.

[0072] The free amino acid contents of 16 kinds of jasmine tea measured in this experiment and the results of one-way ANOVA Figure 1 (B), the results show that there are differences in the free amino acid content of the 16 types of jasmine tea, among which the free amino acid content of sample C is the highest, reaching 0.208%. The free amino acid content of A, B, D, and P teas is similar, close to C, but there are differences. The free amino acid content of F and N teas is close, ranking third, and the amino acid content of J and L teas is also close. The amino acid content is at an intermediate level among all samples. The amino acid content of G, H, K, and M tea samples is low, and that of E and O teas is even lower. The amino acid content of I tea is the lowest among all teas. In summary, the free amino acid content of the 16 types of teas is from high to low: C>A, B, D, P>F, N>J, L>G, H, K, M>E, O>I.

[0073] (3) Tea polyphenols content

[0074] Gallic acid standard curve regression equation: y = 0.01033x + 0.0321, correlation coefficient R 2 =0.99935.

[0075] The tea polyphenols content of 16 kinds of jasmine tea measured in this experiment and the results of one-way ANOVA Figure 1 (C) The results show that there are obvious differences in the content of tea polyphenols among the 16 types of jasmine tea. Among them, the tea polyphenol content of sample M is the highest, significantly higher than that of the other 15 tea samples; there is no difference in the tea polyphenol content between sample E and sample I, both of which are the lowest. The content of tea polyphenols is from high to low: M>L>K>C, G, N>A, B, D>H, P>J, O>F, I.

[0076] (4) Caffeine content

[0077] Caffeine standard curve regression equation: y = 0.05913x + 0.00677, correlation coefficient R 2 =0.99958.

[0078] The free amino acid contents of 16 kinds of jasmine tea measured in this experiment and the results of one-way ANOVA are as follows: Figure 1(D) As shown. The results show that there are obvious differences in the free amino acid content of the 16 types of jasmine tea, among which the free amino acid content of sample C is the highest, significantly higher than the other 15 tea samples; the free amino acid content of sample M is the lowest, significantly lower than the other 15 tea samples. From the figure, it can be seen that the caffeine content of the 16 types of tea is: L>K>C>G, N>H, P>B>A>D>F>E, M, O>I>J.

[0079] Example 2: Correlation analysis between color difference characteristic values ​​of dry jasmine tea and tea polyphenols and caffeine content of tea leaves

[0080] The color difference characteristic values ​​of different jasmine tea dry teas are shown in Table 1. As can be seen from Table 1, the L value of jasmine tea dry tea is in the range of 12.33-42.54, with an average value of 27.45, indicating that the color of its dry tea varies from darker to brighter in brightness. The a value fluctuates in the range of 0.88-13.07, with an average value of 3.74; the b value is in the range of 3.98-24.12, with an average value of 15.43, and the b / a is in the range of 0.58-21.42, with an average value of 9.80. In terms of hue, it is shown that the color of jasmine tea dry tea varies from green to red, and from blue to yellow. △E is in the range of 3.83-75.87, with an average value of 55.97, and the value of color difference △E is large, indicating that the degree of color difference is large. In summary, the color characteristics of jasmine tea dry tea can be described as having a certain brightness change, the color varies from green to red, and from blue to yellow, and there is a large color difference.

[0081] Table 1 Color difference characteristic values ​​of 16 kinds of jasmine tea dry tea

[0082]

[0083]

[0084] The Pearson correlation analysis principle was used to analyze the correlation between the color difference data characteristics and the tea polyphenols and caffeine content of tea. The Origin drawing software was used to draw a correlation heat map. The results are as follows: Figure 2 As shown. Pearson correlation analysis is a commonly used statistical method to measure the linear relationship between two variables. The Pearson correlation coefficient (r) is the covariance divided by the product of the standard deviations of the two variables. When 0<|r|<1, it indicates that there is a certain degree of linear relationship between the two variables. The closer the correlation coefficient between the data is to 1, that is, the red color in the color temperature graph on the right, the stronger the positive correlation between the two; the closer the correlation coefficient between the data is to -1, that is, the blue color in the color temperature graph on the right, the stronger the negative correlation between the two. The correlation coefficient is calculated after statistical analysis of the data. When P<0.05, it means that the two groups of data are significantly correlated, and when P<0.01, it means that the two groups of data are extremely significantly correlated.

[0085] in accordance with Figure 2 It was found that the original tea color difference characteristics a (red and green color components) and △E (total color difference) were significantly correlated with the physical and chemical indicators of jasmine tea. Among them, a was extremely significantly positively correlated with tea polyphenols (P<=0.01), with a correlation coefficient of 0.69; it was significantly correlated with caffeine (P<=0.05), with a correlation coefficient of 0.50. △E was extremely significantly negatively correlated with tea polyphenols in jasmine tea (P<=0.01), with a correlation coefficient of -0.78; it was significantly negatively correlated with caffeine in tea leaves (P<=0.05), with correlation coefficients of -0.57.

[0086] Example 3: Correlation analysis between tea soup image color feature values ​​and tea polyphenols and caffeine content

[0087] Tea soup color is one of the important evaluation factors in tea sensory evaluation. Therefore, the color information of tea soup becomes the main characteristic of tea soup quality. In this study, the HIS color system in image analysis technology is used to express the color characteristics of tea soup, that is, to analyze the H hue, I brightness, and S saturation of each sample. The color feature values ​​of tea soup images of different jasmine samples are shown in Table 2.

[0088] Table 2 HIS mean values ​​of 16 jasmine tea soup images

[0089]

[0090] The Pearson correlation analysis principle was used to analyze the correlation between the color feature data of the tea soup images of different jasmine samples and the content of free amino acids, tea polyphenols, and caffeine. The Origin drawing software was used to draw a correlation heat map. The results are as follows: Figure 3 As shown, the specific parameter explanations in the figure are the same as those in Example 2.

[0091] observe Figure 3 It was found that the mean value of the I (brightness) channel in the HIS channel mean of the tea soup image color was extremely significantly positively correlated with the caffeine content in jasmine tea (P <= 0.01), with a correlation coefficient of 0.64. This may be because the caffeine in tea can be converted into color-enhancing substances. There are reports that different withering temperatures and times of tea will affect the content of caffeine and other substances in tea, thereby affecting the color change of tea soup. This experiment also further illustrates that the caffeine content in tea leaves has an extremely significant correlation with the I brightness of the tea soup color.

[0092] Example 4: Construction and verification of tea polyphenols and caffeine prediction model

[0093] (1) Determination of the optimal number of hidden layer nodes in the neural network model

[0094] The neural network toolbox in MATLAB software randomly divides the sample set into a training set (70%) and a prediction set (30%).

[0095] Tea polyphenols BP neural network model: the input layer is 2, which are the original tea color difference feature a (red and green color components) and △E (total color difference) feature values, the output layer is 1, i.e. tea polyphenols, the hidden layer is 1, and the number of nodes is set to 2-11;

[0096] Caffeine BP neural network model: the input layer is 3, which are the original tea color difference feature a (red and green color components) and △E (total color difference) feature values ​​and the I brightness of the tea soup color; the output layer is 1, i.e. caffeine; the hidden layer is 1, and the number of nodes is set to 2-11.

[0097] The training results are shown in Table 3.

[0098] Table 3 Mean square error (MSE) of the number of hidden layer nodes for each quality index of the training set

[0099]

[0100]

[0101] The number of hidden layer nodes corresponding to the minimum mean square error (MSE) of each quality index in the training set is the best. As shown in Table 3, the optimal number of hidden layer nodes for tea polyphenols and caffeine is 7 and 4 respectively.

[0102] (2) Determination of the prediction model for tea polyphenols content

[0103] The BP prediction model for tea polyphenols content was determined as follows: Tea polyphenols BP neural network model: the input layer is 2, which are the original tea color difference feature a (red and green color components) and △E (total color difference) feature values, the hidden layer is 1, the number of nodes is 7, and the output layer is 1, namely tea polyphenols.

[0104] After the genetic algorithm (GN) optimized BP neural network (GN-BP) model, the MAE, MSE, RMSE, and MAPE indicators were reduced to more than 60% of the original. Therefore, the performance of the tea polyphenols quantitative BP prediction model can be effectively improved based on the genetic algorithm. The results are shown in Table 4.

[0105] Table 4 Performance evaluation indexes of tea polyphenols BP and GN-BP models

[0106]

[0107] The prediction of the BP neural network for the prediction set samples before and after the optimization is observed, as shown in Table 5. It is found that the prediction results of the first three prediction set samples based on the BP model are all lower than the true value, and the relative gap with the true value is large, with the maximum error of -0.0477. The BP model after GN optimization has significantly narrowed the gap between the predicted value and the measured value, among which the gap between the predicted value and the true value of sample 3 has been reduced from the original 0.1649 to 0.0125. The maximum error is -0.0180, and the average absolute error is 4.26%, which is less than 5%, and is statistically reliable.

[0108] Table 5 Prediction error of tea polyphenols BP neural network model before and after optimization

[0109]

[0110] (3) Determination of the prediction model for caffeine content

[0111] Caffeine BP neural network model: the input layer is 3, which are the original tea color difference feature a (red and green color components) and △E (total color difference) feature values ​​and the I brightness of the tea soup color, the hidden layer is 1, the number of nodes is 4, and the output layer is 1, that is, caffeine.

[0112] As shown in Table 6, the performance indicators (MAE, MSE, RMSE, and MAPE) of the caffeine quantitative prediction BP model optimized by GN were all reduced by more than 60% of the original ones, among which the mean square error (MSE) indicator was reduced to 87% of the original one.

[0113] Table 6 Performance evaluation indicators of caffeine BP and GN-BP models

[0114]

[0115]

[0116] The prediction set was tested based on the BP model and the BP prediction model optimized by GN. The results are shown in Table 7. Overall, the error between the predicted value of the BP model and the measured value is large, with the lowest error being 0.00124 and the highest error being -0.0477. The error between the overall predicted value of the BP model optimized by GN and the true value is significantly reduced, with the highest error being 0.0180 and the lowest error being -0.005. The average absolute error is 4.24%, which is less than 5% and is statistically reliable.

[0117] Table 7 Prediction value and error of caffeine BP neural network before and after optimization

[0118]

[0119] Based on the above examples, the time used by this method to detect tea polyphenols and caffeine in jasmine tea can be compared with the national standard chemical analysis method, and the results are shown in Table 8.

[0120] Table 8 Time consumption comparison between the method of the present invention and the traditional national standard method

[0121]

[0122]

[0123] As shown in Table 8, the national standard chemical analysis method takes about 4 hours and 3.5 hours to detect jasmine tea polyphenols and caffeine, respectively. The two indicators are measured for a total of about 7.5 hours, and the time is distributed in multiple operation links such as sample pretreatment, test reagent preparation, drawing standard curve, sample reaction time, and sample detection time. The time used for the two indicators of the method of the present invention is close, both about 35 minutes, and the two indicators can be detected at the same time. The time is only distributed in the two operation links of sample pretreatment and detection. No chemical reagents are needed, which is green and environmentally friendly and easy to operate. The time consumption accounts for only 7.76% of the time used by the traditional national standard detection method, which greatly improves the detection efficiency.

[0124] The method does not use chemical reagents, is environmentally friendly, can simultaneously detect two indicators of jasmine tea, tea polyphenols and caffeine, is time-saving, easy to operate, and can be applied to on-site rapid detection of jasmine tea.

[0125] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principles of the present invention shall be equivalent replacement methods and shall be included in the protection scope of the present invention.

Claims

1. A method for rapid simultaneous detection of tea polyphenols and caffeine content in jasmine tea, characterized in that The method is: Firstly, the color difference characteristic values ​​of the dry tea samples were measured using a colorimeter; Then, the tea soup was brewed with boiling water and put into a self-made "image acquisition dark box" to collect the tea soup image, and the tea soup image feature value was extracted using MATLAB; The acquired dry tea color difference and tea soup image feature values ​​are input into the constructed BP neural network prediction model to calculate the tea polyphenols and caffeine contents in the samples respectively.

2. The method according to claim 1, characterized in that The method comprises the following steps: (1) Evenly spread a certain amount of dried jasmine tea samples on a white cardboard, pay attention to keep the samples even, avoid direct light exposure to the samples, use a colorimeter to measure at least one of the sample's L brightness, a green-red intensity, b blue-yellow intensity, and ΔE total color difference, each sample is measured three times in parallel, and the average value is taken; (2) Accurately weigh 2-10 g of jasmine tea sample, brew it in boiling water at a ratio of 1:50-80 g / g, let it stand for 3-10 min, put the prepared tea soup in a white porcelain bowl, cool it to about 20-50° C., put it in an "image acquisition dark box" to take pictures, and use MATLAB software to extract relevant feature values ​​of the tea soup image, wherein the relevant feature values ​​are one or more of H hue, I brightness, and S saturation color. (3) The obtained relevant dry tea color difference and tea soup image feature values ​​are respectively input into the constructed BP neural network prediction model of jasmine tea polyphenols and caffeine to calculate the content of tea polyphenols and caffeine in the sample.

3. The method according to claim 2, characterized in that In step (1), a colorimeter is used to measure the green-redness and ΔE total color difference characteristic value of the sample.

4. The method according to claim 2, characterized in that In step (2), 3 g of tea sample is weighed, brewed in boiling water at a ratio of 1:50 g / ml, and allowed to stand for 5 minutes. The sample is placed in a self-made "enclosed image acquisition box" containing an LED downlight, and photographed by a smartphone with a pixel of 48 million or more.

5. The method according to claim 2, characterized in that As described in step (2), after collecting the tea soup image, the image is denoised by using the median filtering method, the image is enhanced by using the histogram equalization method, and the color system characteristic values ​​of the tea soup image H hue, I brightness, and S saturation are extracted by using a program edited by MATLAB software.

6. The method according to claim 2, characterized in that The BP neural network prediction model is a BP neural network model optimized by genetic algorithm GN.

7. The method according to claim 2, characterized in that As described in step (3), the obtained a, △E eigenvalues ​​of the color difference of dry tea and the I brightness eigenvalue of the tea soup image are respectively input into the constructed BP neural network prediction model of jasmine tea polyphenols and caffeine, and the tea soup image is collected, the image is denoised by the median filtering method, the image is enhanced by the histogram equalization method, and the program edited by MATLAB software is used to extract the H hue, I brightness, and S saturation color system eigenvalues ​​of the tea soup image.

8. The method according to claim 7, characterized in that As described in step (3), the BP prediction model for tea polyphenol content in step (3) is determined as: tea polyphenol BP neural network model: the input layer is 2, which are the original tea color difference characteristics a green-red intensity and △E total color difference characteristic value, the hidden layer is 1, the number of nodes is 7, and the output layer is 1.

9. The method according to claim 7, characterized in that In the step (3), the caffeine BP neural network model has: an input layer of 3, which respectively includes the original tea color difference feature a green-red intensity and the △E total color difference feature value and the I brightness of the tea soup color; a hidden layer of 1, a node number of 4, and an output layer of 1.

Images