Rapid detection method for flavonoid components and moisture in persimmon leaves and products thereof

Abstract

The invention discloses a method for rapidly detecting flavonoid components and moisture in persimmon leaves and products thereof. The method comprises the following steps: measuring the content of six flavonoid components and the content of moisture in a calibration set sample; scanning the persimmon leaf sample by using a near infrared spectrometer to obtain near infrared spectrum data; importing the obtained near infrared spectrum data of the sample into Brucker infrared spectrometer OPUS 6.5 software, analyzing the spectrum information, establishing a near infrared spectrum model by taking the spectrum information which is not preprocessed as a contrast and combining a partial least square method, and constructing a correction model of the relationship between the six flavonoid components and the moisture content of the persimmon leaf sample and the characteristic near infrared spectrum data; near infrared spectrum data of persimmon leaves to be detected are collected according to spectrum collection parameters the same as those of samples in the calibration set, the six flavonoid components and the moisture content in the persimmon leaves to be detected are obtained according to the constructed calibration model, the detection work efficiency is greatly improved, and comprehensive control over the product quality is facilitated.

Description

Technical Field

[0001] This invention relates to a rapid detection method for flavonoids and moisture in persimmon leaves and their products. Background Technology

[0002] Persimmon leaves are the dried leaves of the persimmon (Diospyros kaki Thunb.), a plant in the Ebenaceae family. Current research shows that persimmon leaves mainly contain flavonoids, terpenoids, tannins, coumarins, and other components. The content determination methods mostly employ high-performance liquid chromatography (HPLC), thin-layer scanning, or colorimetric methods to detect the content of one or more components such as oleanolic acid, ursolic acid, quercetin, kaempferol, quercetin-7-O-β-D-rhamnoside, genistein, and total flavonoids. Among these methods, HPLC has the highest accuracy. Some scholars have used a single-test, multiple-evaluation method to simultaneously determine six flavonoid components, including hyperoside, isoquercitrin, quercetin, myricetin, quercetin, and kaempferol, in Naoxinqing tablets (a traditional Chinese medicine made from persimmon leaf extract); or they have used high-performance liquid chromatography-mass spectrometry to simultaneously determine eight active ingredients, including protocatechuic acid, hyperoside, kaempferol-3-O-β-D-glucopyranoside, kaempferol-3-O-β-D-galactopyranoside, myricetin, quercetin, naringenin, and kaempferol, in Naoxinqing raw material extract and its tablets; Yin Renjie et al. used HPLC-DAD to construct the fingerprint spectrum of Naoxinqing tablets and found 16 common peaks in the common pattern chromatogram to help comprehensively control its quality.

[0003] The existing quality control technologies for persimmon leaf medicinal materials and their products mostly involve detecting a single component or simultaneously determining 2 to 6 components in the persimmon leaf. Most analytical methods for simultaneously determining the content of multiple components are high performance liquid chromatography, or thin layer chromatography to determine the content of triterpenoid components, and colorimetric methods to determine the content of total flavonoids. All of the above detection methods require a series of operations before detection, such as crushing or grinding the sample, sonication, heating and boiling, alcohol precipitation or extraction. The sample pretreatment time is long and the operation is complicated.

[0004] Near-infrared spectroscopy (NIRS) is a high-energy vibrational spectroscopy technique with wavelengths between 780 nm and 2526 nm. This technique, combined with chemometric algorithms, constructs a model detection method for qualitative or quantitative analysis of unknown samples through the establishment and validation of calibration models. It features fast analysis speed, no use of toxic or harmful reagents, no damage to samples, and simultaneous analysis of multiple components in samples. It can realize qualitative and quantitative analysis and online real-time determination of traditional Chinese medicine and its preparations. It is simple to operate and can improve detection efficiency and reduce analysis costs.

[0005] There are currently no reports on the use of near-infrared spectroscopy to detect flavonoids and moisture in persimmon leaves and their products. Summary of the Invention

[0006] The purpose of this invention is to provide a rapid method for detecting flavonoids and moisture in persimmon leaves and their products.

[0007] This invention is achieved through the following technical solutions:

[0008] A rapid method for detecting flavonoids and moisture in persimmon leaves and their products, comprising the following steps:

[0009] 1) Collect calibration set samples and validation set samples from persimmon leaf samples;

[0010] 2) The contents of six flavonoid components and the moisture content of the persimmon leaf calibration set samples were determined. The six flavonoid components were hyperoside, isoquercitrin, astragaloside, myricetin, and quercetin.

[0011] 3) Scan the persimmon leaf samples with a near-infrared spectrometer to obtain near-infrared spectral data;

[0012] 4) Import the near-infrared spectral data of the sample obtained in step 3) into the Brucker infrared spectrometer OPUS 6.5 software, analyze the spectral information, and preprocess the data using the first derivative (FD) method. Using the unprocessed spectral information as a control, a near-infrared spectral model was established using partial least squares method. A calibration model was constructed to establish the relationship between the content and moisture content of six flavonoid components in persimmon leaf samples (hyperoside, isoquercitrin, astragaloside, myricetin, and quercetin) and the characteristic near-infrared spectral data.

[0013] 5) Near-infrared spectral data of the persimmon leaves to be tested were collected using the same spectral acquisition parameters as the calibration set samples. Based on the calibration model constructed in step 4), the six flavonoid components and moisture content in the persimmon leaves to be tested were obtained.

[0014] Preferably, the persimmon leaf sample is pulverized and the powder is passed through a No. 3 sieve.

[0015] The number of calibration set samples is 15 or more, preferably 46.

[0016] The beneficial effects of this invention are as follows: The sample can be measured immediately after pulverization, eliminating the need for complex pretreatment operations such as heating and extraction, making it simple and quick. This invention can rapidly and simultaneously detect the moisture content and the content of six chemical components (hyperoside, isoquercitrin, astragaloside, myricetin, quercetin, and kaempferol) in persimmon leaves. This detection method greatly improves the efficiency of detection work and is beneficial to the comprehensive control of product quality. Attached Figure Description

[0017] Figure 1 This is the original near-infrared reflectance spectrum of the persimmon leaf calibration set sample in Example 1.

[0018] Figure 2 This is the near-infrared spectrum of the persimmon leaf calibration set sample preprocessed by the first derivative method in Example 1.

[0019] Figure 3 This is an external validation of the near-infrared prediction model for the content of six flavonoid components and moisture in persimmon leaves in Example 1.

[0020] Figure 4 This is an external validation of the near-infrared prediction model for the content of six flavonoid components and moisture in persimmon leaf tea in Example 2. Detailed Implementation

[0021] The following is a further description of the invention, but not a limitation thereof.

[0022] Instruments and Reagents: Agilent 1260 High Performance Liquid Chromatograph (Agilent Technologies, USA). Agilent Eclipse Plus C18 column (250 mm × 4.6 mm, 5 µm). Bruker MPA Near Infrared Spectrometer (BRUKER GmbH, Germany). JA3003 Electronic Analytical Balance (Shanghai Balance Instrument Factory), SB25-12DTD Ultrasonic Cleaner (Ningbo Xinzhi Biotechnology Co., Ltd.)

[0023] Hyperoside (batch number MUST-16102605), isoquercetin (MUST-17051005), astragaloside (MUST-16120510), myricetin (MUST-17022504), and quercetin (MUST-16111114) were all purchased from Chengdu Mansite Biotechnology Co., Ltd., and kaempferol (batch number 110861-201611) was purchased from the National Institutes for Food and Drug Control. Acetonitrile (chromatographic grade, Tedia, USA), formic acid (analytical grade, Guangzhou Chemical Reagent Factory), and distilled water were used in the experiments. All other reagents were of analytical grade.

[0024] Samples: 60 batches of persimmon leaves were collected from Henan, Hebei, Shandong, Shaanxi and other producing areas from September 2020 to December 2021. They were identified by Dr. Zhang Huiye of Guangzhou Baiyunshan Hutchison Whampoa Chinese Medicine Co., Ltd. as dried leaves of the persimmon plant (Diospyros kaki Thunb.). The details of each batch of persimmon leaf samples are shown in Table 1.

[0025] Table 1 Information on Persimmon Leaf Samples

[0026] Table 1 Diospyros kaki leaf sample information table

[0027] The 30 batches of persimmon leaf tea samples were developed by Guangzhou Baiyunshan Hutchison Whampoa Pharmaceutical Co., Ltd.

[0028] Example 1:

[0029] 1. 46 out of 60 batches of persimmon leaf samples were set as calibration set samples for near-infrared spectroscopy detection, and 14 batches were set as validation set samples.

[0030] II. The content of six flavonoids in 46 batches of persimmon leaf calibration set samples was determined by a single-test-multiple-evaluation method. The flavonoid content was referenced from the HPLC detection method for six flavonoids in persimmon leaf extract established by Guo Haibiao et al. (Guo Haibiao, Dong Fuyue, Li Wenshan, et al. Determination of six flavonoids in persimmon leaf extract by a single-test-multiple-evaluation method [J]. Chinese Journal of Modern Applied Pharmacy, 2021, 38(7): 831-835). Chromatographic conditions: The chromatographic column was an Agilent Eclipse Plus C18 (250 mm × 4.6 mm, 5 μm); the mobile phase was acetonitrile (A)-0.1% phosphoric acid solution (B); the gradient elution conditions were 0-40 min 7%-15% A, 40-60 min 75%-50% A; the flow rate was 1.0 ml / min; the injection volume was 10 μl; the column temperature was 25℃; and the detection wavelength was 256 nm.

[0031] Preparation of the mixed reference solution: Accurately weigh appropriate amounts of each reference standard (hyperoside, isoquercetin, astragaloside, myricetin, quercetin, and kaempferol), place them in a 25 mL volumetric flask, dissolve and dilute to the mark with methanol to obtain the mixed reference stock solution. Accurately measure an appropriate amount of the mixed reference stock solution, dilute to the mark with methanol, and prepare a solution containing 100 μg / mL of each component. -1 The mixed reference standard stock solution.

[0032] Preparation of the test solution: Weigh 0.1 g of persimmon leaf powder (passed through a No. 3 sieve) accurately, place it in a 50 mL stoppered conical flask, add 20 mL of methanol accurately, stopper tightly, weigh, sonicate (250 W, 45 kHz) for 30 min, remove, cool to room temperature, replenish the lost mass with methanol, shake well, filter through a 0.45 μm microporous membrane, and collect the filtrate to obtain the test solution.

[0033] The contents of six flavonoids in persimmon leaves were determined using the external standard two-point method: the test solution was analyzed under the chromatographic conditions described above, and the contents of the six flavonoids in persimmon leaves were calculated using the external standard two-point method. The average contents of the six components in 46 batches of calibration set samples were 648.51 μg / g, 1211.63 μg / g, 1540.43 μg / g, 1629.38 μg / g, 121.24 μg / g, and 162.82 μg / g, respectively; the average contents of the six components in 14 batches of validation set samples were 669.86 μg / g, 1246.69 μg / g, 1527.58 μg / g, 1725.35 μg / g, 129.69 μg / g, and 182.91 μg / g, respectively.

[0034] Determination of moisture content in persimmon leaves: Approximately 5g of persimmon leaf powder was weighed and the moisture content of 60 batches of persimmon leaf samples was determined according to Method II (drying method) of "0832 Moisture Determination" in Part IV of the 2020 edition of the Chinese Pharmacopoeia. The average moisture content of 46 batches of calibration set samples was 9.05%, and the average moisture content of 14 batches of validation set samples was 9.74%.

[0035] Table 2. Flavonoid composition and moisture content of persimmon leaf samples in the calibration and validation sets (n) 校正集 =46, n 验证集 =14)

[0036] III. Establishment of a near-infrared prediction model for the content of persimmon leaf flavonoids

[0037] 1. Near-infrared spectral information acquisition

[0038] Near-infrared spectral information of persimmon leaf powder samples (passed through a No. 3 sieve) was acquired using a Bruker MPA near-infrared spectrometer. The spectral range of this near-infrared spectrometer is 780 ~ 2526 nm, or 3960 cm⁻¹. -1 ~ 12800 cm -1 The resolution was 5 nm. Before collecting spectral information, the persimmon leaf powder sample to be scanned was placed in the laboratory where the near-infrared spectrometer was located for 24 hours to adapt to environmental factors such as temperature and reduce their impact on the experimental process. The near-infrared spectrometer was then preheated for 30 minutes before the sample powder was scanned and the spectral information was collected. The experiment was repeated 3 times, with each sample loading repeated twice.

[0039] 2. Preprocessing and Modeling of Near-Infrared Spectral Information

[0040] The spectral information acquired by the near-infrared spectrometer was imported into the Brucker infrared spectrometer OPUS 6.5 software for analysis. Preprocessing methods included standard normal variable transformation (SNV), minimum-maximum normalization (MMN), multivariate scattering correction (MSC), first derivative (FD), second derivative (SD), FD+SNV, and FD+MSC. Using unprocessed spectral information as a control, a near-infrared spectral model was established using partial least squares method.

[0041] The spectra and component contents of the persimmon leaf calibration set samples were imported into the Brucker infrared spectrometer OPUS 6.5 software for preprocessing and optimization. The obtained root mean square error of calibration set (RMSEc), correlation coefficient of calibration set (Rc), root mean square error of cross-validation set (RMSEv), and correlation coefficient of cross-validation set (Rv) were used to compare the eight spectral preprocessing optimization methods. The smaller the RMSEc and RMSEv, and the closer Rc and Rv are to 1, the better the prediction effect of the model.

[0042] The near-infrared spectral preprocessing results corresponding to hyperoside are shown in Table 3. The optimal spectral preprocessing method for the near-infrared prediction model of persimmon hyperoside is min-max normalization. The prediction model established by this method has 9 principal components and a spectral range of 6102-4597.7 cm⁻¹. -1 The Rc value was 0.9314, the RMSEc value was 98.6, the Rv value was 0.8854, and the RMSEv value was 120, indicating that the model had good predictive performance. The near-infrared spectral preprocessing results corresponding to isoquercitrin are shown in Table 4. The optimal spectral preprocessing method for the near-infrared prediction model of isoquercitrin in persimmon leaves is the second derivative. The prediction model established by this method has 5 principal components and a spectral range of 7502.1-5446.3 cm⁻¹. -1 The Rc value was 0.9325, the RMSEc value was 190, the Rv value was 0.8122, and the RMSEv value was 295, indicating that the model has good predictive performance. The near-infrared spectral preprocessing results corresponding to astragalin are shown in Table 5. The optimal spectral preprocessing method for the persimmon leaf astragalin near-infrared prediction model is the first derivative. The prediction model established by this method has 9 principal components and a spectral range of 7502.1-5446.3 cm⁻¹. -1 The Rc value was 0.9456, the RMSEc value was 168, the Rv value was 0.8505, and the RMSEv value was 258, indicating that the model has good predictive performance. The near-infrared spectral preprocessing results corresponding to myricetin are shown in Table 6. The optimal spectral preprocessing method for the near-infrared prediction model of myricetin in persimmon leaves is the second derivative. The prediction model established by this method has 10 principal components and a spectral range of 6102-4597.7 cm⁻¹. -1The Rc values ​​were 0.9787, RMSEc was 127, Rv was 0.9234, and RMSEv was 225, indicating that the model had good predictive performance. The near-infrared spectral preprocessing results corresponding to quercetin are shown in Table 7. The optimal spectral preprocessing method for the near-infrared prediction model of persimmon leaf quercetin was the standard normal variable transformation method. The prediction model established by this method had 10 principal components and a spectral range of 12493.3-6098.1 cm⁻¹. -1 The Rc value was 0.9906, the RMSEc value was 14.6, the Rv value was 0.9241, and the RMSEv value was 38.7, indicating that the model has good predictive performance. The near-infrared spectral preprocessing results corresponding to kaempferol are shown in Table 8. The optimal spectral preprocessing method for the near-infrared prediction model of kaempferol from persimmon leaves is min-max normalization. The prediction model established by this method has 10 principal components and a spectral range of 12493.3-4246.7 cm⁻¹. -1 The Rc value was 0.9822, the RMSEc value was 27.1, the Rv value was 0.931, and the RMSEv value was 50, indicating that the model has good predictive performance. The near-infrared spectral preprocessing results corresponding to moisture content are shown in Table 9. The optimal spectral preprocessing method for the near-infrared prediction model of persimmon leaf moisture is no preprocessing. The prediction model established by this method has 10 principal components and a spectral range of 6102-4246.7 cm⁻¹. -1 The Rc values ​​are 0.9988, RMSEc is 0.12, Rv is 0.9978, and RMSEv is 0.157, indicating that the model has good predictive performance.

[0043] Table 3. Effects of different near-infrared spectral processing methods on hyperoside prediction model parameters.

[0044] Table 4. Effects of different near-infrared spectral processing methods on the parameters of the isoquercitrin prediction model.

[0045] Table 5. Effects of different near-infrared spectral processing methods on the parameters of the astragaloside prediction model.

[0046] Table 6. Effects of different near-infrared spectral processing methods on the parameters of the myricetin prediction model.

[0047] Table 7. Effects of different near-infrared spectral processing methods on quercetin prediction model parameters

[0048] Table 8. Effects of different near-infrared spectral processing methods on the parameters of the kaempferol prediction model.

[0049] Table 9. Effects of different near-infrared spectral processing methods on parameters of moisture prediction models.

[0050] Because simultaneous quantification of seven components in the sample was required, and the optimal preprocessing method for the near-infrared spectra of each component differed, the first derivative (FD) method was ultimately selected as the optimal preprocessing method for all seven components after a comprehensive comparison of the preprocessing results for each component. Under this method, the calibration set correlation coefficients (Rc) for hyperoside, isoquercetin, astragaloside, myricetin, quercetin, kaempferol, and water were all greater than 0.92, and the cross-validation set correlation coefficients (Rv) were all greater than 0.82. The unprocessed near-infrared spectra (…) Figure 1 ) and the near-infrared spectrum after first-order derivative preprocessing ( Figure 2 By comparison, it was found that the spectral curves after preprocessing were more concentrated and the absorption peaks were more obvious.

[0051] IV. Validation and Evaluation of Near-Infrared Prediction Model for Flavonoid Content in Persimmon Leaves

[0052] To verify the accuracy of the near-infrared prediction model, the near-infrared spectra of 14 samples from the persimmon leaf validation set were imported into the software, and the near-infrared prediction model established by the calibration set was used for prediction. The predicted values ​​were compared with the actual values ​​using correlation coefficient. The closer the correlation coefficient is to 1, the closer the predicted value is to the measured value.

[0053] The correlations between predicted and measured values ​​of various components in 14 (n=2) persimmon leaf samples in the validation set were as follows: hyperoside R=0.9955, isoquercetin R=0.9461, astragaloside R=0.9144, myricetin R=0.9052, quercetin R=0.905, kaempferol R=0.9033, and moisture R=0.9982. The correlations between predicted and measured values ​​for each component were all greater than 0.9, indicating significant correlations. These results demonstrate the high accuracy of the model established by this method. The external validation results of the near-infrared prediction model for the content of six flavonoid components and moisture content in persimmon leaves are shown below. Figure 3 As shown (i.e., the correlation results between predicted and measured values).

[0054] Example 2

[0055] 1. 22 out of 30 batches of persimmon leaf tea samples were set as calibration set samples for near-infrared spectroscopy detection, and 8 batches were set as validation set samples.

[0056] II. The content of six flavonoid components in 30 batches of persimmon leaf calibration set samples was determined using a single-test-multiple-evaluation method. Referring to step two of Example 1, the average content of the six components in 22 batches of calibration set samples were 638.02 μg / g, 1174.11 μg / g, 1432.03 μg / g, 1489.38 μg / g, 112.56 μg / g, and 155.37 μg / g, respectively; and the average content of the six components in 8 batches of validation set samples were 696.04 μg / g, 1345.49 μg / g, 1827.31 μg / g, 2098.35 μg / g, 152.51 μg / g, and 200.91 μg / g, respectively.

[0057] Determination of moisture content in persimmon leaves: Approximately 5g of persimmon leaf powder was weighed and the moisture content of 30 batches of persimmon leaf samples was determined according to Method II (drying method) of "0832 Moisture Determination" in Part IV of the 2020 edition of the Chinese Pharmacopoeia. The average moisture content of 22 batches of calibration set samples was 9.23%, and the average moisture content of 8 batches of validation set samples was 9.18%.

[0058] III. Establishment of a near-infrared prediction model for flavonoid content in persimmon leaf tea samples

[0059] 1. Near-infrared spectral information acquisition

[0060] Same as Example 1

[0061] 2. Preprocessing and Modeling of Near-Infrared Spectral Information

[0062] Referring to Example 1, the near-infrared spectral preprocessing results corresponding to hyperoside are shown in Table 10. The optimal spectral preprocessing method for the near-infrared prediction model of persimmon hyperoside is min-max normalization. The prediction model established by this method has 5 principal components and a spectral range of 12493.3-4246.7 cm⁻¹. -1 The Rc values ​​were 0.8932, RMSEc was 134, Rv was 0.801, and RMSEv was 159, indicating that the model had good predictive performance. The near-infrared spectral preprocessing results for isoquercitrin are shown in Table 11. The optimal spectral preprocessing method for the near-infrared prediction model of isoquercitrin in persimmon leaves was FD+SNV. The prediction model established by this method had 7 principal components and a spectral range of 7502.1-4597.7 cm⁻¹. -1 The Rc value was 0.9283, the RMSEc value was 202, the Rv value was 0.7912, and the RMSEv value was 290, indicating that the model has good predictive performance. The near-infrared spectral preprocessing results corresponding to astragalin are shown in Table 12. The optimal spectral preprocessing method for the persimmon leaf astragalin near-infrared prediction model is the standard normal variable transformation method. The prediction model established by this method has 2 principal components and a spectral range of 12493.3-4246.7 cm⁻¹. -1The Rc value was 0.8022, the RMSEc value was 301, the Rv value was 0.7398, and the RMSEv value was 320, indicating that the model has good predictive performance. The near-infrared spectral preprocessing results corresponding to myricetin are shown in Table 13. The optimal spectral preprocessing method for the near-infrared prediction model of myricetin in persimmon leaves is the standard normal variable transformation method. The prediction model established by this method has 8 principal components and a spectral range of 7502.1-4246.7 cm⁻¹. -1 The Rc value was 0.9686, the RMSEc value was 219, the Rv value was 0.8911, and the RMSEv value was 332, indicating that the model has good predictive performance. The near-infrared spectral preprocessing results corresponding to quercetin are shown in Table 14. The optimal spectral preprocessing method for the near-infrared prediction model of persimmon leaf quercetin is min-max normalization. The prediction model established by this method has 10 principal components and a spectral range of 7502.1-4246.7 cm⁻¹. -1 The Rc values ​​were 0.9794, RMSEc was 27.8, Rv was 0.8335, and RMSEv was 59.7, indicating that the model had good predictive performance. The near-infrared spectral preprocessing results for kaempferol are shown in Table 15. The optimal spectral preprocessing method for the persimmon leaf kaempferol near-infrared prediction model was the first derivative. The prediction model established by this method had 5 principal components, a spectral range of 12493.3-4597.7 cm⁻¹, an Rc value of 0.9834, an RMSEc value of 27.3, an Rv value of 0.7588, and an RMSEv value of 87.6, indicating that the model had good predictive performance. The near-infrared spectral preprocessing results for moisture are shown in Table 16. The optimal spectral preprocessing method for the persimmon leaf moisture near-infrared prediction model was the first derivative. The prediction model established by this method had 7 principal components, a spectral range of 6102-4597.7 cm⁻¹. -1 The Rc values ​​are 0.9982, RMSEc is 0.157, Rv is 0.9954, and RMSEv is 0.214, indicating that the model has good predictive performance.

[0063] Table 10. Effects of different near-infrared spectral processing methods on hyperoside prediction model parameters

[0064] Table 11. Effects of different near-infrared spectral processing methods on the parameters of the isoquercitrin prediction model.

[0065] Table 12 Effects of different near-infrared spectral processing methods on the parameters of the astragaloside prediction model.

[0066] model

[0067] Table 13. Effects of different near-infrared spectral processing methods on the parameters of the myricetin prediction model.

[0068] model

[0069] Table 14. Effects of different near-infrared spectral processing methods on quercetin prediction model parameters.

[0070] Table 15. Effects of different near-infrared spectral processing methods on the parameters of the kaempferol prediction model.

[0071] Table 16. Effects of different near-infrared spectral processing methods on parameters of moisture prediction models.

[0072] 3. Validation and evaluation of the near-infrared prediction model for flavonoid content in persimmon leaf tea

[0073] If simultaneous quantification of seven components in a sample is required, the prediction model needs to have high stability. However, due to the current calibration set of 22 samples, the established model has low stability. Furthermore, the optimal preprocessing method for the near-infrared spectra of each component is inconsistent. After repeated testing, it was found that the external validation correlation for simultaneous quantification of seven components is too low (i.e., correlation coefficient < 0.7). Therefore, this embodiment chooses to establish prediction models for each of the seven components separately and perform external validation.

[0074] To further verify the accuracy of the near-infrared prediction model, the near-infrared spectra of eight persimmon leaf samples from the validation set were imported into the software, and the near-infrared prediction model established using the calibration set was used for prediction. The predicted values ​​were compared with the actual values. The external validation results of the near-infrared prediction model for the content of six flavonoid components and moisture in persimmon leaves are as follows: Figure 4 As shown in the figure, the correlations between the predicted and measured values ​​of each component in the eight persimmon leaf samples in the validation set were as follows: hyperoside R=0.9087, isoquercetin R=0.9257, astragaloside R=0.9275, myricetin R=0.9808, quercetin R=0.9139, kaempferol R=0.9403, and moisture R=0.9978. The results indicate that the correlations between the predicted and measured values ​​of each component are all greater than 0.9, which is significant, indicating that the established model has high prediction accuracy.

[0075] The results of this experiment show that it is difficult to establish an effective multi-component near-infrared prediction model when the number of samples is very limited. In this case, it is advisable to establish a model for each component separately and verify it.

Claims

1. A method for rapid detection of flavonoid components and moisture in persimmon leaves and products thereof, characterized in that, The method includes the following steps: 1) Collect calibration set samples and validation set samples from persimmon leaf samples; 2) The contents of six flavonoid components and the moisture content of the persimmon leaf calibration set samples were determined. The six flavonoid components were hyperoside, isoquercitrin, astragaloside, myricetin, and quercetin. 3) Scan the persimmon leaf samples with a near-infrared spectrometer to obtain near-infrared spectral data; 4) Import the near-infrared spectral data of the sample obtained in step 3) into the Brucker infrared spectrometer OPUS 6.5 software, analyze the spectral information, use the first derivative as the preprocessing method, use the unprocessed spectral information as a control, and combine partial least squares method to establish a near-infrared spectral model to construct a correction model of the relationship between the content and moisture content of the six flavonoid components of persimmon leaf sample: hyperoside, isoquercitrin, astragaloside, myricetin, and quercetin and the characteristic near-infrared spectral data. 5) Near-infrared spectral data of the persimmon leaves to be tested were collected using the same spectral acquisition parameters as the calibration set samples. Based on the calibration model constructed in step 4), the six flavonoid components and moisture content in the persimmon leaves to be tested were obtained.

2. The method according to claim 1, characterized in that, The persimmon leaf sample was pulverized and the powder was passed through a No. 3 sieve and a sieve with a aperture smaller than No.

3.

3. The method according to claim 1, characterized in that, The spectral range of the near infrared spectrometer is 7502.1 ~ 4246.7 cm -1 .

4. The method according to claim 1, characterized in that, The number of samples in the calibration set is 15 or more.

5. The method according to claim 1, characterized in that, The calibration set contains 46 samples.

Images