KR20220063897A - Method for identifying age of ginseng and processed ginseng using NIRS - Google Patents

Abstract

The present invention relates to an age determination method of a 4-year-old, 5-year-old, and 6-year-old ginseng and processed ginseng using a near infrared reflectance spectroscopy (NIRS). The age determination method of a ginseng and processed ginseng using a near infrared reflectance spectroscopy is a method for rapidly and accurately determining a ginseng age, in which an absorption spectrum of a sample for a specific range is simply collected and analyzed without separate and complicated preprocessing for the ginseng and processed ginseng that are hard to be determined by naked eyes. Therefore, the age determination method is useful for accurately determining an age of a ginseng and processed ginseng on the market. The age determination method comprises the following steps of: measuring an infrared absorption spectrum; comparing and statistically analyzing differences in the spectrums for each ginseng age; and determining an age of one or more ginseng sample by using the statistical analysis and an age determination prediction model.

Description

{Method for identifying age of ginseng and processed ginseng using NIRS}

The present invention relates to a method for determining lotus roots for ginseng and processed ginseng using a Near Infrared Reflectance Spectroscopy (NIRS). It relates to a method for discriminating the lotus root of ginseng and processed ginseng through the calibration equation for determining the lotus root of ginseng and processed ginseng prepared by MPLS (Modified Partial Least Squares) regression analysis method.

Ginseng ( Panax ginseng CA Meyer) is a perennial herbaceous plant belonging to the genus of ginseng classified in the family Araliaceae or Araliaceae.

Ginseng is divided into fresh ginseng harvested from the field and not dried according to the processing type; white ginseng, which is dried without the fine roots removed by sunlight or hot air, and red ginseng, dried by steaming fresh ginseng with steam. are doing

In addition, the lotus root of ginseng refers to the number of years that ginseng sprouted and grown. From a long time ago, there have been reports of differences in the amount of various physiologically active ingredients and physiologically active ingredients according to the lotus root of ginseng. Among them, a significant difference has occurred in the saponin content. Since there is a significant difference, it is very important to identify the lotus root of ginseng.

In foreign countries, the identification of lotus root is not significant because ginseng is harvested and distributed after 3 or 4 years of age. However, in Korea, consumers think that the closer the 6-year-old is, the better the medicinal effect and efficacy. In this expensive distribution market, there are frequent cases of false labeling of 4-year roots as 5- or 6-year roots and 5-year roots as 6-year roots, which disrupts the distribution order and lowers consumer confidence. It is emerging as a very important issue.

According to the current Ministry of Agriculture, Food and Rural Affairs Ordinance of the Ministry of Agriculture, Food and Rural Affairs to determine the lotus root, the inspection staff attends and judges the lotus root at harvest time. It is required to be identified (Ministry of Agriculture, Food and Rural Affairs, Ginseng Industry Act, Act No. 16101, 2018), which is used as an auxiliary method of identification in the field where the cultivation history is confirmed, such as the planting year, but it is controversial because it is only a sensory evaluation level, In the distribution market (direct transaction, wholesale, retail, etc.) where it is difficult to confirm the cultivation history, it is difficult to apply because the reliability is not high.

As for the lotus root identification method of ginseng introduced in the literature, the expert's visual discrimination by stem trace confirmation (Non-Patent Document 1), the expert's visual discrimination by the brain head trace confirmation (Non-Patent Document 2), and histochemical staining of the ginseng secretion duct Although the method of lotus root identification (Non-Patent Document 3) by am. However, since the tissue staining method of ginseng secretion duct is a method to discriminate by the root section of the main root, it is limited to fresh ginseng and is difficult to apply to processed ginseng, that is, white ginseng, red ginseng, and Taegeuk ginseng.

In addition, a study on age discrimination according to ginsenoside content using HPLC-MS (Non-Patent Document 4), a study on the ginsenoside content of leaves and roots of ginseng according to age (Non-Patent Document 5), and HPLC-DAD Changes in physiologically active substances due to the age and seasonal changes of used ginseng (Non-Patent Document 6) and a study on the metabolite analysis technology of ginseng using HPLC-qTOF-MS (Non-Patent Document 7).

In addition, the method for determining the lotus root of ginseng using a chromatography mass spectrometer (Patent Document 1), the method for determining the lotus root of ginseng using 1H-NMR and multivariate statistical analysis (Patent Document 2), and the method for determining the origin and lotus root of ginseng roots (Patent Document) 3), a standard marker for determining the origin or age of processed ginseng, a method for establishing it, or a method for determining the origin or age using the same (Patent Document 4) and a biomarker composition for discriminating ginseng lotus root (Patent Document 5) However, most of these are indicators showing differences between groups by extracting primary or secondary metabolites with organic solvents using physicochemical analysis equipment such as LC-MS, GC-MS, qTOF, and NMR and analyzing them using multivariate analysis. It relates to a method for determining the origin or age of ginseng or processed ginseng by detecting a substance and establishing it as a standard biomarker.

In the preceding study, pretreatment of a rather complicated and difficult extraction process using organic solvents is essential, and through this, the lotus root from ginseng and processed ginseng was distinguished by the difference in components such as ginseng saponin. The invention of the accurate lotus root identification technology of ginseng and processed ginseng using an absorption spectrum in a specific near-infrared region using a near-infrared spectrometer that can be analyzed is not currently taught or known.

On the other hand, near-infrared spectroscopy analyzes the spectrum reflected or transmitted by irradiating a sample with near-infrared rays in the 400-2,500 nm region using a near-infrared spectrometer. It is characterized by absorption bands of overtones and combinations of fundamental frequencies, and the absorption energy intensity is relatively weak. This analysis method was not used for actual analysis before the development of chemical statistical processing techniques, but it was used for analysis of crude protein, moisture, crude fat, etc. has been applied in various fields.

This near-infrared spectroscopy method reduces the effort, time, and expense required for analysis, does not require the use of toxic reagents used in general chemical analysis, etc. are being developed, and it is widely known as a method suitable for characterization of food, feed, and livestock products in a moisture-containing state.

Accordingly, in the present invention, as a result of diligent efforts to quickly and accurately determine the lotus root of ginseng and processed ginseng, a method for determining the lotus root of ginseng and processed ginseng using a near-infrared spectrometer was developed. The absorption spectrum of a specific area of a sample was collected and analyzed by freeze-drying or simple grinding of commercial ginseng and processed ginseng lotus root suspected of having lotus root without pretreatment, confirming that it was possible to quickly and accurately discriminating the lotus root, and completed the present invention.

Registered Patent No. 10-1149236 (2012.05.16) Registered Patent No. 10-1202282 (2012.11.12) Registered Patent No. 10-1273574 (2013.06.04) Registered Patent No. 10-1409848 (2014.06.13) Registered Patent No. 10-1444579 (2014.09.18)

Lee, J. H. et al., J. Ginseng Res. 20(1):72-77, 1996 Lee, J. H. et al., J. Ginseng Res. 31(3):142-146, 2007 Lee, J. H. et al., J. Ginseng Res. 25(2):101-105, 2001 Wang Y. et al., J. Phytochem. Anal. 17:424-430, 2006 Li X. et al., J. Life Sci. 9:670-683, 2012 Shan SM. et al., J. Pharm. Biomed. Anal. 89:76-82, 2014 Mao Q. et al., J. Pharm. Biomed. Anal. 97:129-140, 2014

The present invention was conceived to solve the above problems, and the present invention analyzes the rapid and accurate identification of lotus root of ginseng and processed ginseng using a near-infrared spectrometer without a separate complicated pre-processing process, and commercially available ginseng with suspected lotus root And an object of the present invention is to provide a method for determining the lotus root by collecting the absorption spectrum of a specific region of processed ginseng.

In order to achieve the above object, the present invention

(a) measuring the near-infrared absorption spectrum for each wavelength by irradiating one or more ginseng samples with near-infrared rays;

(b) confirming the difference in the spectrum for each lotus root of the near-infrared absorption spectrum obtained for the one or more ginseng samples, and performing statistical analysis;

(c) determining the lotus root of one or more ginseng samples to be discriminated using the statistical analysis and the lotus root discrimination prediction model; provides a method for determining lotus root of ginseng and processed ginseng, comprising:

In one embodiment of the present invention, the lotus root discrimination prediction model of step (c) is

(c-1) measuring a near-infrared absorption spectrum for each wavelength by irradiating a plurality of ginseng samples with near-infrared rays;

(c-2) arbitrarily dividing the plurality of ginseng samples for development of a prediction model for determining a lotus root and for verifying a prediction model for determining a lotus root;

(c-3) Obtain a raw absorption spectrum obtained by averaging the near-infrared absorption spectra obtained for the plurality of ginseng samples, its first derivative or second derivative, respectively, and check the peak difference according to wavelength in the spectrum for each lotus root and developing a predictive model for determining lotus root by performing statistical analysis;

(c-4) comparing the lotus root discrimination prediction model with the near-infrared absorption spectrum for lotus root discrimination prediction model verification, performing statistical analysis to verify the fit of the lotus root discrimination prediction model, and selecting the lotus root discrimination prediction model; including may be selected.

In one embodiment of the present invention, the ginseng sample is a fresh ginseng sample or a processed ginseng sample, and the lotus root of the fresh ginseng or processed ginseng may be a 4-year, 5-year, or 6-year root. The shape of ginseng less than 3 years old is clearly distinguished with the naked eye during distribution, and the product value is low, so there is little need to distinguish it separately because the yield or distribution amount is very small. hard There is a large difference in the distribution price, so it is necessary to determine precisely.

In one embodiment of the present invention, the processed ginseng may be one or more selected from the group consisting of raw ginseng, dried ginseng, white ginseng, gok ginseng, jik ginseng, and red ginseng.

In one embodiment of the present invention, the ginseng sample may be a main root or a fine root.

In one embodiment of the present invention, the ginseng sample may be prepared by freeze-drying to a moisture content of 8% or less. During freeze-drying, effective long-term storage is possible by removing moisture while maintaining the tissue state of ginseng (fresh ginseng) samples with high moisture content.

In one embodiment of the present invention, the near-infrared absorption spectrum of step (a) may be to use an absorption spectrum in a wavelength band of 400 to 2,500 nm.

According to a preferred embodiment of the present invention, it is preferable to measure the absorption spectrum between 400 and 2,500 nm for the measurement of the near-infrared absorption spectrum for the development of the prediction model for determining the lotus root of ginseng and the processed ginseng and the spectrum collection for the verification of the prediction model. In general, the near-infrared absorption spectrum uses an absorption spectrum between 700 and 2,500 nm, but the absorption spectral pattern is different depending on the main components constituting the sample of ginseng and processed ginseng, providing important information for the discrimination analysis of lotus root of ginseng and processed ginseng. Therefore, it can be said that it is undesirable to take an absorption spectrum with a limited wavelength of 700 to 2,500 nm.

In one embodiment of the present invention, the ratio of the number of ginseng samples for predictive model development or ginseng samples for predictive model verification when randomly dividing one or more ginseng samples in step (b) may be 1:0.2 to 1. .

The measured near-infrared absorption spectra for each wavelength of ginseng and processed ginseng are arbitrarily divided for development of a prediction model for lotus root discrimination and for verification of a prediction model. When the ratio of the number of ginseng samples is less than 1:0.2, the selected prediction models are applied to the actual unknown sample. It is difficult to verify whether the accuracy is applied high even when it is used, and if it exceeds 1:1, it may be difficult to develop a large number of predictive models suitable for identification of fresh ginseng lotus root using absorption spectra of multiple samples.

In one embodiment of the present invention, the statistical analysis in step (c) or (d) is multiple linear regression (MLR), partial least squares (PLS), modified partial least squares method. (MPLS: Modified Partial Least Squares) and PCA: Principle Component Analysis (PCA) may be performed by one or more methods selected from the group consisting of.

When multivariate analysis methods such as partial least squares method, modified partial least squares method and component analysis method are applied as the method applied to the above statistical analysis, the physicochemical information of the analyte sample is not substantially investigated according to the physicochemical analysis method, There is an advantage of statistically and rapidly simultaneously analyzing the variables of the analysis sample included in a specific population, and the application of this multivariate analysis method is a quantitative analysis method (Patent Documents 1 to 5, Non-Patent Documents 1 to 7) through the various pretreatment methods introduced above. ) shows a significant difference when compared with

In addition, another reason for performing various statistical analyzes is the overlapping of absorption wavelengths generated in the near-infrared absorption spectrum. When developing a prediction model, in most cases, absorption wavelengths overlap, reducing the accuracy of prediction. In the present invention, it is possible to provide an excellent predictive model that provides a more precise prediction while excluding the influence of overlapping wavelengths through the statistical analysis.

In one embodiment of the present invention, the method for determining the difference for each lotus root in step (c) is a raw absorption spectrum obtained by correcting an average of the near-infrared absorption spectrum obtained for one or more ginseng samples, its first derivative derivative, or 2 It may be to obtain a differential derivative, respectively, and to check the peak difference according to wavelength in the spectrum for each lotus root.

The average corrected raw absorption spectrum is specifically corrected by averaging the near-infrared absorption spectrum arbitrarily classified for the development of the lotus root discrimination prediction model with the number of ginseng and processed ginseng used in the analysis. Using the average far-infrared absorption spectrum as the far-infrared absorption spectrum, applying its derivative to confirm the difference in the spectrum for each lotus root, and performing various statistical analyzes, it is possible to provide an optimal lotus root discrimination prediction model of ginseng and processed ginseng. It is possible.

The application of the first derivative derivative or the second derivative derivative is because the absorbance shift phenomenon occurs in the measured far-infrared absorption spectrum due to various factors. When water treatment (math treatment) is performed by applying the first or second derivative to solve the movement phenomenon, all near-infrared absorption spectra are aligned to the same baseline, so the difference in absorption power derived from the movement phenomenon is eliminated, and specific absorption Since discrimination power is improved using only the difference in absorption of wavelengths, discrimination and discrimination of specific spectra are made more clearly through derivative processing. This has the advantage of increasing field applicability by excluding factors that interfere with the analysis, such as the absorption shift of the spectrum, easily, making it easy to identify lotus roots, even for the general public who are not specially trained for the analysis.

In one embodiment of the present invention, it may be to verify the accuracy of the predictive model by using the modified partial least squares discriminative analysis predictive model to which the first derivative is applied in the discriminative predictive model in step (d).

This is to analyze whether the developed lotus root discrimination predictive model of ginseng and processed ginseng is accurately applied to unknown samples, and to select the optimal predictive model by measuring the precision at that time.

In one embodiment of the present invention, the statistical analysis may be to discriminate the lotus root of the ginseng sample according to the three-dimensional distribution position of each sample in the three-dimensional stereoscopic discriminant power test in the lotus root discrimination prediction model, which is a principal component analysis method.

Most of the 4-year-old ginseng samples are distributed in the center of the 1st, 2nd, and 3D planes, and the 5-year-old or 6-year-old ginseng samples are clearly separated and distributed, so it is clearly a group of 4-year-old, 5-year-old and 6-year-old ginseng samples. It is possible to discriminate between

In one embodiment of the present invention, the statistical analysis is distributed in the range of 3.37 to 4.49 random variables in the case of a 4-year-old ginseng sample, and a random variable in the case of a 5-year-old ginseng sample in the lotus root discrimination prediction model, where the statistical analysis is a modified partial least-squares discriminant analysis method. It is distributed in the range of 4.50-5.49, and in the case of a 6-year-old ginseng sample, it may be distributed in a random variable of 5.50-6.65.

The arbitrary variable range may vary depending on the type of ginseng or processed ginseng (red ginseng, white ginseng).

The present invention is a technology for accurately and quickly discriminating the lotus root of ginseng and processed ginseng using a near-infrared spectrometer. Since it is possible to quickly and accurately identify lotus roots by collecting and analyzing absorption spectra in a specific region of a sample of ginseng, it has the effect of being useful for identification of lotus roots of ginseng and processed ginseng sold in the market.

The effect of the present invention is to provide a technique for accurately and quickly discriminating the lotus root of ginseng and processed ginseng using a near-infrared spectrometer.
For commercially available ginseng and processed ginseng, which is difficult to track history and identify lotus root, in reality, it is possible to identify lotus root in a relatively simple way without the need for a separate complicated pre-treatment process or extraction with an organic solvent.
In addition, it is possible to quickly identify by collecting and analyzing absorption spectra of commercially available ginseng and processed ginseng samples suspected of having lotus roots. can be utilized.
1 is an average raw near-infrared absorption spectrum of 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples obtained in Example 1.
2 is a near-infrared absorption spectrum of the average first derivative of 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples obtained in Example 1. FIG.
3 is a near-infrared absorption spectrum of the average second derivative of 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples obtained in Example 1.
4 is a diagram showing three-dimensional stereoscopic discrimination patterns according to principal component analysis of near-infrared absorption spectra of 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples obtained in Example 1. [+: 4-year root, □: 5-year root, ◇ : 6 years old].
5 is a diagram showing the discriminant pattern according to the modified partial least-squares analysis method (first derivative) of the near-infrared absorption spectrum of the 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples obtained in Example 1. FIG.
6 is a diagram showing the discriminating pattern of unknown samples according to the modified partial least-squares analysis method (first derivative) of the near-infrared absorption spectra of 4-year-old, 5-year-old and 6-year-old fresh ginseng samples obtained in Example 1. FIG.
7 is an average raw near-infrared absorption spectrum of 4-year-old, 5-year-old, and 6-year-old white ginseng samples obtained in Example 2.
8 is a near-infrared absorption spectrum of the average first derivative of 4-year-old, 5-year-old, and 6-year-old white ginseng samples obtained in Example 2.
9 is a near-infrared absorption spectrum of the average second derivative of 4-year-old, 5-year-old, and 6-year-old white ginseng samples obtained in Example 2.
10 is a diagram showing the three-dimensional stereoscopic discrimination pattern according to the principal component analysis of the near-infrared absorption spectrum of the 4-year-old, 5-year-old and 6-year-old white ginseng samples obtained in Example 2 [+: 4-year root, □: 5-year root, ◇ : 6 years old].
11 is a diagram showing the discrimination patterns according to the modified partial least squares analysis method (first derivative) of the near-infrared absorption spectra of 4-year-old, 5-year-old, and 6-year-old white ginseng samples obtained in Example 2. FIG.
12 is a diagram showing the discriminating pattern of unknown samples according to the modified partial least-squares analysis method (first derivative) of the near-infrared absorption spectra of 4-year-old, 5-year-old and 6-year-old white ginseng samples obtained in Example 2. FIG.
13 is an average far-infrared absorption spectrum of 4-year-old, 5-year-old, and 6-year-old red ginseng samples obtained in Example 3.
14 is a near-infrared absorption spectrum of the average first derivative of the 4-year-old, 5-year-old, and 6-year-old red ginseng samples obtained in Example 3.
15 is a near-infrared absorption spectrum of the average second derivative of the 4-year-old, 5-year-old, and 6-year-old red ginseng samples obtained in Example 3.
16 is a diagram showing three-dimensional stereoscopic discrimination patterns according to principal component analysis of the near-infrared absorption spectrum of 4-year-old, 5-year-old and 6-year-old red ginseng samples obtained in Example 3 [+: 4-year-old root, □: 5-year-old root, ◇ : 6 years old].
17 is a diagram showing the discriminant pattern according to the modified partial least squares analysis method (first derivative) of the near-infrared absorption spectrum of the 4-year-old, 5-year-old and 6-year-old red ginseng samples obtained in Example 3. FIG.
18 is a diagram showing the discriminating pattern of unknown samples according to the modified partial least squares analysis method (first derivative) of the near-infrared absorption spectrum of the 4-year-old, 5-year-old, and 6-year-old red ginseng samples obtained in Example 3. FIG.

The specific content of the present invention will be described in detail with reference to examples. The following examples are only for illustrating the present invention, and the scope of the present invention is not limited by the following examples and drawings.

[Example 1. Development of prediction model for identification of lotus root of fresh ginseng]

1) Predictive model development

In the present invention, in order to perform accurate lotus root identification of fresh ginseng using a near-infrared spectrometer, the National Agricultural Products Quality Management Institute directly visited a farm (producer) and secured 98 4-year-old fresh ginseng, 101 5-year-old root, and 122 6-year-old root. The equipment used to develop the predictive model was the NIRS XDS Model (Foss, U.S.A.). Fresh ginseng by lotus root obtained from various regions in Korea is washed thoroughly with running tap water, rinsed with ultrapure water at least twice, then sliced at 2-3 mm intervals with a ceramic knife, then frozen at -40°C for 12 hours and then frozen with a freeze dryer for 72 hours. It was dried (moisture content less than 8%). After freeze-drying, pulverize the sample with a micro hammer mill (Automill, tokken, Japan) and pass it through a 1 mm sieve to keep it refrigerated (around 4℃) in a plastic bag with a zipper to prevent moisture absorption until analysis. After allowing time, it was used for analysis.

In order to collect the spectrum of fresh ginseng samples for each lotus root using a near-infrared spectrometer, 5 g of the stored sample was placed in a small cup and the absorbance for each wavelength between 400 and 2,500 nm was measured to obtain a raw near-infrared absorption spectrum, Each of the measured samples, 98 4-year-old fresh ginseng, 101 5-year-old fresh ginseng, and 122 6-year-old fresh ginseng samples were randomly divided into two groups.

That is, one group of 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples was used as a sample for developing a predictive model for lotus root discrimination using 60 4-year-old, 69 5-year-old, and 83 6-year-old fresh ginseng samples. The lotus root discrimination predictive model ( calibration) was evaluated.

As a result of averaging the raw near-infrared absorption spectra of 212 samples for developing a prediction model for lotus root discrimination of fresh ginseng, some wavelengths such as 780, 942, 1002, 1464, 1874, 1938 and 2004 nm among the entire wavelength range from 400 to 2,500 nm There was a difference in the near-infrared absorption spectra of 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples.

In addition, as shown in Fig. 2, the near-infrared absorption spectrum that appears when the average raw spectrum of the 4-year, 5-year, and 6-year-old fresh ginseng samples measured at a wavelength of 400 to 2,500 nm is applied to the first and second derivatives when applied to the first derivative The average spectrum of the 4-year-old fresh ginseng sample showed slightly lower absorbance than the spectrum of the 5-year-old and 6-year-old fresh ginseng samples at 898, 972, and 1904 nm, whereas the average spectrum of the 6-year-old fresh ginseng sample was at 1418 and 2038 nm at 1418 and 2038 nm. It showed slightly lower absorption characteristics than the fresh ginseng samples, so it was possible to discriminate between 4-year-old, 5-year-old and 6-year-old fresh ginseng samples by the difference in absorption spectra in a specific wavelength region. In addition, as shown in FIG. 3, even when the second derivative is applied, the average spectrum of 4-year, 5-year, and 6-year-old fresh ginseng samples are absorption spectra at 884, 914, 984, 1196, 1332, 1364, 1404, 1436, 1890 and 1928 nm. , and the difference in absorption spectrum occurred in a similar manner to the result of the first derivative, making it possible to discriminate the lotus roots of 4-year, 5-year, and 6-year-old fresh ginseng.

And, as a result of examining the three-dimensional discrimination power of 4-year-old, 5-year-old, and 6-year-old fresh ginseng through Principle Component Analysis (PCA) using the near-infrared absorption spectrum, 4-year root, 4-year-old root, The difference between 5-year-old and 6-year-old fresh ginseng samples was clearly seen, and the groups of 4-year-old fresh ginseng samples were mostly distributed in the 1st, 2nd, and 3D planes, but the 5-year-old fresh ginseng samples moved to the left and rear in the 2D plane. , 6-year-old fresh ginseng samples were distributed in the front left, and it was confirmed that the 4-year-old, 5-year-old, and 6-year-old fresh ginseng samples were clearly distinguishable between groups.

In addition, when the derivative was applied to the raw absorption spectrum of the fresh ginseng sample for each lotus root, the modified partial least squares discriminant analysis (MPLSDA), which is one of the multivariate statistical analysis methods of the near-infrared spectral spectrum, was used for 4-year roots, 5 Table 1 below shows the results of examining the discriminant power prediction models of the fresh ginseng samples of year-old and 6-year-old roots.

Math treatment N Calibration cross-validation SEC R ² Bias SECV RSC SD 0,0,1,1 212 0.225 0.924 0.003 0.351 2.291 0.804 1,4,4,1 212 0.175 0.956 0.023 0.290 2.848 0.826 2,8,8,1 212 0.197 0.945 0.018 0.266 3.147 0.837 N: samples used to develop the model
SEC: standard error of calibration
R ² : coefficient of determination of calibration
Bias: average difference between reference and NIRS values
SECV: standard error of cross-validation
RSC: SD/SECV, the ratio of SD (standard deviation) to SECV in the calibration set

As shown in Table 1 above, when the variables to be discriminated are set to 4, 5-year-old fresh ginseng group, 5, and 6-year-old fresh ginseng group by 3 classification methods, modified partial least squares discrimination analysis method As a result of the analysis ^of fresh ginseng lotus root discrimination according to , in the case of 5-year-old fresh ginseng, random variables 4.59~5.49, and in the case of 6-year-old fresh ginseng, random variables 5.50~6.40, it was confirmed that the scattering distribution was accurately classified. It was confirmed that the lotus root discrimination of fresh ginseng samples was 100% possible, and the results are shown in Table 2 below.

division correct discriminant number
(n) incorrect discriminant number
(n) accuracy
(%) 4 years old
(60 types) Primitive Derivatives (0,0,1,1) 60 0 100 First derivative(1,4,4,1) 60 0 100 Second derivative(2,8,8,1) 60 0 100 5 years old
(69 types) Primitive Derivatives (0,0,1,1) 69 0 100 First derivative(1,4,4,1) 69 0 100 Second derivative(2,8,8,1) 69 0 100 6 years old
(83 types) Primitive Derivatives (0,0,1,1) 83 0 100 First derivative(1,4,4,1) 83 0 100 Second derivative(2,8,8,1) 83 0 100

In the above results, when 212 types of fresh ginseng were classified into 4-year, 5-year, and 6-year roots with different cultivated lotus roots, the raw spectrum, the first and second derivative spectra of each lotus root, and the first and second derivative spectra and a statistical multivariate analysis method. The result was obtained that it is possible to discriminate the lotus root of fresh ginseng by three-dimensional three-dimensional analysis through principal component analysis and the least-squares discriminating analysis method for deformed parts.

2) Predictive model verification

On the other hand, when the prediction model developed using the modified partial least-squares discriminant analysis method is applied to the daily analysis for identification of lotus roots of fresh ginseng samples, 38 types of 4-year-old fresh ginseng can be used as a predictive model for lotus root identification. Table 3 below shows the results of verifying the accuracy when applied to unknown samples of , 32 5-year-old roots, and 39 6-year-old roots.

Math treatment SD Bias r ² SEP(C) slope RSP 0,0,1,1 0.783 0.025 0.892 0.278 1.018 2.817 1,4,4,1 0.846 0.016 0.920 0.242 0.957 3.496 2,8,8,1 0.839 0.037 0.895 0.276 0.952 3.040 Mean: average of used sample values
SD: standard deviation of mean
Bias: average difference between reference and NIRS values
r ² : coefficient of determination of cross-validation
SEP(C): the corrected standard error of prediction
RSP: SD/SEP(C), the ratio of SD of reference data to SEP(C) in the external validation set

As can be seen from the results of Table 3 above, the modified partial least squares discriminative analysis predictive model developed for the discrimination of fresh ginseng lotus root to which the first derivative is applied has a coefficient of determination (r ² ) of 0.920, and the coefficient of determination when developing a predictive model It showed a similar level of discrimination accuracy, and as can be seen in Table 4 below, in the case of 4-year-old fresh ginseng, the random variable was distributed in the range of 3.46 to 4.48, and the discrimination accuracy was 97.4%, and in the case of 5-year-old fresh ginseng, the random variable was in the range of 4.52 to 5.52. The discriminant accuracy was 96.9% by distribution, and in the case of 6-year-old fresh ginseng, it was distributed in the random variable range of 5.46 to 6.38, so the discrimination accuracy was 97.4%. It was confirmed again that fresh ginseng lotus root can be identified at a possible level (see FIG. 6).

division correct discriminant number
(n) incorrect discriminant number
(n) accuracy
(%) 4 years old
(38 types) Primitive Derivatives (0,0,1,1) 37 One 97.4 First derivative(1,4,4,1) 37 One 97.4 Second derivative(2,8,8,1) 36 2 94.7 5 years old
(32 types) Primitive Derivatives (0,0,1,1) 32 0 100.0 First derivative(1,4,4,1) 31 One 96.9 Second derivative(2,8,8,1) 31 One 96.9 6 years old
(39 types) Primitive Derivatives (0,0,1,1) 39 0 100.0 First derivative(1,4,4,1) 38 One 97.4 Second derivative(2,8,8,1) 36 3 92.3

[Example 2. Development of predictive model for determining lotus root of white ginseng]

1) Predictive model development

In the present invention, in order to perform accurate lotus root identification of white ginseng using a near-infrared spectrometer, 130 kinds of 4-year-old white ginseng, 124 kinds of 5-year-old root, and 103 6-year-old roots were obtained. The equipment used to develop the predictive model was the NIRS XDS Model (Foss, U.S.A.).

White ginseng for each lotus root obtained from various regions in Korea is cut with a medicinal herb cutter at 3-5 mm intervals, then pulverized with a grinder (Pulverisette 19, FRITSCH, Germany), passed through a 1 mm sieve, and refrigerated in a zippered plastic bag to prevent moisture absorption until analysis. After storage (within 4℃), if necessary, take it out and leave it at room temperature for 12 hours, and then use it for analysis.

To collect the spectrum of white ginseng samples for each lotus root using a near-infrared spectrometer, 5 g of the stored sample was placed in a small cup and the absorbance for each wavelength between 400 and 2,500 nm was measured to obtain a raw near-infrared absorption spectrum, Each of the measured samples, 130 kinds of 4-year-old white ginseng, 124 kinds of 5-year-old white ginseng, and 103 kinds of 6-year-old white ginseng were randomly divided into two groups.

That is, one group of 4-year-old, 5-year-old, and 6-year-old white ginseng samples was used as a sample for developing a prediction model for lotus root discrimination using 86 4-year-old white ginseng, 99 5-year-old root, and 53 6-year-old white ginseng samples. The lotus root discrimination prediction model developed using 44 kinds of 4-year-old white ginseng, 25 kinds of 5-year-old root, and 50 kinds of 6-year-old root discrimination predictive model developed using as a verification sample (for predictive model validation) for the evaluation of unknown samples. calibration) was evaluated.

As a result of averaging the raw near-infrared absorption spectra of 238 samples for the development of a prediction model for lotus root discrimination of white ginseng, as shown in FIG. 7, 542-770, 1130, 1320, 1466, 1678, 1862, 1934, 2018 among the entire wavelength range of 400 to 2,500 nm. And it was confirmed that there is a clear difference in the near-infrared absorption characteristics when the lotus root of white ginseng is different by showing differences in the near-infrared absorption spectra of 4-year-old, 5-year-old and 6-year-old white ginseng samples in some wavelength ranges such as 2096 nm.

In addition, as shown in FIG. 8, the near-infrared absorption spectrum that appears when the average raw spectrum of the 4-year-old, 5-year-old, and 6-year-old white ginseng samples measured at a wavelength of 400 to 2,500 nm is applied to the first and second derivatives when applied to the first derivative The average spectrum of the 4-year-old white ginseng sample showed slightly lower absorbance than that of the 5-year-old and 6-year-old white ginseng samples at 606~758, 786, and 1902 nm, whereas the average spectrum of the 6-year-old white ginseng sample was 1352, 1416, and 2046 nm, etc. It showed slightly lower absorption characteristics than the 4-year-old and 5-year-old white ginseng samples, so it was possible to discriminate the 4-year-old, 5-year-old and 6-year-old white ginseng samples by the difference in the absorption spectrum in a specific wavelength region.

In addition, as shown in FIG. 9, even when the second derivative is applied, the average spectra of 4-year-old, 5-year-old, and 6-year-old white ginseng samples are 504~526, 588, 782, 916, 986, 1362, 1398, 1436, 1702, 1886, 1924. , 2016, 2066-2106, and 2236 nm showed differences in absorption spectra, and differences in absorption spectra occurred in a similar manner to the results of the first derivative, making it possible to discriminate the lotus roots of 4-year-old, 5-year-old and 6-year-old white ginseng.

And, as a result of examining the three-dimensional discrimination power of 4-year-old, 5-year-old, and 6-year-old white ginseng through Principle Component Analysis (PCA) using the near-infrared absorption spectrum, as shown in FIG. The difference between 5-year-old and 6-year-old white ginseng samples was clearly revealed, and most of the groups of 4-year-old white ginseng samples were distributed in the center of the primary, secondary, and 3D planes, but the 5-year-old white ginseng samples moved to the right and rear in the 2D plane. , 6-year-old white ginseng samples were distributed in the front left corner, so it was possible to clearly identify the 4-year-old, 5-year-old and 6-year-old white ginseng samples between groups.

In addition, when the derivative was applied to the raw absorption spectrum of the white ginseng sample for each lotus root, the modified partial least squares discriminant analysis (MPLSDA), which is one of the multivariate statistical analysis methods of the near-infrared spectral spectrum, was used for 4-year root, 5 Table 5 below shows the results of examining the discriminant power prediction models of the two-year-old and six-year-old white ginseng samples.

Math treatment N Calibration cross-validation SEC R ² Bias SECV RSC SD 0,0,1,1 238 0.227 0.909 0.003 0.326 2.236 0.729 1,4,4,1 238 0.191 0.936 0.001 0.322 2.264 0.729 2,8,8,1 238 0.197 0.932 0.001 0.307 2.391 0.734 N: samples used to develop the model
SEC: standard error of calibration
R ² : coefficient of determination of calibration
Bias: average difference between reference and NIRS values
SECV: standard error of cross-validation
RSC: SD/SECV, the ratio of SD (standard deviation) to SECV in the calibration set

As shown in Table 5 above, when the variables to be discriminated are set to 4, 5-year-old white ginseng group, 5, and 6-year-old white ginseng group by 3 classification methods, the modified partial least squares discrimination analysis method As a result of the analysis ^of white ginseng lotus root discrimination according to , in the case of 5-year-old white ginseng, random variables 4.52-5.46, and in the case of 6-year-old white ginseng, the random variables 5.56-6.25 were accurately classified and distributed and the distribution was confirmed. It was confirmed that 100% of lotus root discrimination of white ginseng samples was possible, and the results are shown in Table 6 below.

division correct discriminant number
(n) incorrect discriminant number
(n) accuracy
(%) 4 years old
(86 types) Primitive Derivatives (0,0,1,1) 86 0 100 First derivative(1,4,4,1) 86 0 100 Second derivative(2,8,8,1) 86 0 100 5 years old
(99 types) Primitive Derivatives (0,0,1,1) 99 0 100 First derivative(1,4,4,1) 99 0 100 Second derivative(2,8,8,1) 99 0 100 6 years old
(53 types) Primitive Derivatives (0,0,1,1) 53 0 100 First derivative(1,4,4,1) 53 0 100 Second derivative(2,8,8,1) 53 0 100

In the above results, when 238 kinds of white ginseng were divided into 4-year, 5-year, and 6-year roots by different cultivated lotus roots, the raw spectrum, the first and second derivative spectra of each lotus root, and the first and second derivative spectra and a statistical multivariate analysis method. The result was obtained that it is possible to discriminate the lotus root of white ginseng by three-dimensional three-dimensional analysis through principal component analysis and the least-squares discrimination analysis method for deformed parts.

2) Predictive model verification

On the other hand, when the prediction model developed using the modified partial least squares discrimination analysis method is applied to the daily analysis for discrimination of lotus roots of white ginseng samples, 44 kinds of 4-year-old white ginseng can be used as a predictive model for lotus root identification. Table 7 below shows the results of verifying the accuracy when applied to unknown samples of , 25 5-year-old roots, and 50 6-year-old roots.

Math treatment SD Bias r ² SEP(C) slope RSP 0,0,1,1 0.803 0.035 0.900 0.285 1.053 2.818 1,4,4,1 0.818 0.007 0.923 0.249 1.047 3.285 2,8,8,1 0.819 0.005 0.923 0.249 1.045 3.289 Mean: average of used sample values
SD: standard deviation of mean
Bias: average difference between reference and NIRS values
r ² : coefficient of determination of cross-validation
SEP(C): the corrected standard error of prediction
RSP: SD/SEP(C), the ratio of SD of reference data to SEP(C) in the external validation set

As can be seen from the results of Table 7, the modified partial least squares discriminant analysis predictive model developed for the discrimination of white ginseng lotus root to which the first derivative is applied has a coefficient of determination (r ² ) of 0.923, which is the coefficient of determination when developing the predictive model and It exhibited a similar level of discrimination accuracy, and as can be seen in Table 8 below, in the case of 4-year-old white ginseng, the random variable was distributed in the range of 3.52 to 4.58, and the discrimination accuracy was 97.7%, and in the case of 5-year-old white ginseng, the random variable was in the range of 4.61 to 5.52. In the case of white ginseng root distribution, the discriminant accuracy was 96.0%, and in the case of 6-year-old white ginseng, the random variable was distributed in the range of 5.38 to 6.30, resulting in a discrimination accuracy of 98.0%. It was confirmed again that white ginseng lotus root can be discriminated to a level that is possible (see FIG. 12).

division correct discriminant number
(n) incorrect discriminant number
(n) accuracy
(%) 4 years old
(44 types) Primitive Derivatives (0,0,1,1) 41 3 93.2 First derivative(1,4,4,1) 43 One 97.7 Second derivative(2,8,8,1) 43 One 97.7 5 years old
(25 types) Primitive Derivatives (0,0,1,1) 24 One 96.0 First derivative(1,4,4,1) 24 One 96.0 Second derivative(2,8,8,1) 25 0 100.0 6 years old
(50 types) Primitive Derivatives (0,0,1,1) 49 One 98.0 First derivative(1,4,4,1) 49 One 98.0 Second derivative(2,8,8,1) 49 One 98.0

[Example 3. Development of predictive model for determining lotus root of red ginseng]

1) Predictive model development

In the present invention, in order to perform accurate lotus root identification of red ginseng using a near-infrared spectrometer, 99 kinds of 4-year-old red ginseng, 67 kinds of 5-year-old root, and 45 6-year-old roots were obtained. The equipment used to develop the predictive model was the NIRS XDS Model (Foss, U.S.A.).

Red ginseng by lotus root obtained from various regions in Korea is cut with a medicinal herb cutter at 3-5 mm intervals, then crushed with a grinder (Pulverisette 19, FRITSCH, Germany), passed through a 1 mm sieve, and stored in a plastic bag with a zipper to prevent moisture absorption until analysis. After refrigeration (around 4°C), take it out if necessary, leave it at room temperature for 12 hours, and then use it for analysis.

In order to collect the spectrum of the red ginseng sample for each lotus root using a near-infrared spectrometer, 5 g of the stored sample was placed in a small cup and the absorbance for each wavelength between 400 and 2,500 nm was measured to obtain a raw near-infrared absorption spectrum, Each of the measured samples, 99 kinds of 4-year-old red ginseng, 67 kinds of 5-year-old red ginseng, and 45 kinds of 6-year-old red ginseng were randomly divided into two groups.

That is, one group of each 4-year-old, 5-year-old, and 6-year-old red ginseng sample was used as a sample for developing a predictive model for lotus root discrimination using 77 4-year-old red ginseng, 42 5-year-old, and 28 6-year-old red ginseng samples, and the other group The lotus root discrimination predictive model ( calibration) was evaluated.

As a result of averaging the raw near-infrared absorption spectra of 147 samples for the development of a prediction model for lotus root discrimination of red ginseng, 524 to 780, 1132, 1320, 1462, 1862, 1936, 2018, 2094 in the wavelength range of 400 to 2,500 nm as shown in FIG. And in some wavelength ranges such as 2230 nm, there was a difference in the near-infrared absorption spectra of 4-year-old, 5-year-old, and 6-year-old red ginseng samples.

In addition, as shown in FIG. 14, the near-infrared absorption spectrum that appears when the average raw spectrum of the 4-year-old, 5-year-old, and 6-year-old red ginseng samples measured at a wavelength of 400 to 2,500 nm is applied to the first and second derivatives when applied to the first derivative The average spectrum of the 4-year-old red ginseng sample clearly showed a difference from the spectrum of the 5-year-old and 6-year-old red ginseng samples at 476~512, 552, 786, 970, 1416, 1902, and 1962 nm, whereas the average spectrum of the 6-year-old red ginseng sample showed slightly lower absorption characteristics than the spectrum of 4-year-old and 5-year-old red ginseng samples at 970, 1004-1044, and 1502 nm, making it possible to discriminate 4-year-old, 5-year-old and 6-year-old red ginseng samples by the difference in absorption spectra in a specific wavelength region. And, as shown in FIG. 15, even when the second derivative is applied, the average spectra of 4-year-old, 5-year-old and 6-year-old red ginseng samples are 522, 560-576, 762, 780, 914, 952, 986, 1196, 1360, 1396, 1434. , 1884, 1922~2016, 2066 and 2276 nm showed differences in absorption spectra, and differences in absorption spectra occurred in a similar manner to the results of the first derivative, making it possible to discriminate the lotus roots of 4-year-old, 5-year-old, and 6-year-old red ginseng. .

In addition, as a result of examining the three-dimensional discrimination power of 4-year-old, 5-year-old and 6-year-old red ginseng through Principle Component Analysis (PCA) using the near-infrared absorption spectrum, 4-year root, 4-year-old root, The difference between 5-year-old and 6-year-old red ginseng samples was clearly revealed, and the groups of 4-year-old red ginseng samples were mostly distributed in the 1st, 2nd, and 3D planes, but the 5-year-old red ginseng samples were on the right side in the 2D plane, The 6-year-old red ginseng samples were distributed at the back of the center, so it was possible to clearly identify the 4-year-old, 5-year-old and 6-year-old red ginseng samples between groups.

In addition, when the derivative is applied to the raw absorption spectrum of the red ginseng sample for each lotus root, the modified partial least squares discriminant analysis (MPLSDA), which is one of the multivariate statistical analysis methods of the near-infrared spectral spectrum, was used for 4-year roots, 5 Table 9 shows the results of examining the discriminant power prediction model of the year-old and 6-year-old red ginseng samples.

Math treatment N Calibration cross-validation SEC R ² Bias SECV RSC SD 0,0,1,1 147 0.242 0.906 0.009 0.403 1.943 0.783 1,4,4,1 147 0.216 0.923 0.007 0.407 1.877 0.764 2,8,8,1 147 0.220 0.922 0.015 0.343 2.289 0.785 N: samples used to develop the model
SEC: standard error of calibration
R ² : coefficient of determination of calibration
Bias: average difference between reference and NIRS values
SECV: standard error of cross-validation
RSC: SD/SECV, the ratio of SD (standard deviation) to SECV in the calibration set

As shown in Table 9 above, when the variables to be discriminated are set to 4, 5-year-old red ginseng group, 5, and 6-year-old red ginseng group by three classification methods, the modified partial least squares discrimination analysis method As a result of the analysis ^of red ginseng lotus root discrimination according to , in the case of 5-year-old red ginseng, the random variable 4.62~5.36, and in the case of 6-year-old red ginseng, the random variable 5.50~6.43, it was confirmed that the scattering distribution was accurately classified. It was confirmed that the lotus root discrimination of the red ginseng sample was 100% possible, and the results are shown in Table 10 below.

division correct discriminant number
(n) incorrect discriminant number
(n) accuracy
(%) 4 years old
(77 types) Primitive Derivatives (0,0,1,1) 77 0 100 First derivative(1,4,4,1) 77 0 100 Second derivative(2,8,8,1) 77 0 100 5 years old
(42 types) Primitive Derivatives (0,0,1,1) 41 One 97.6 First derivative(1,4,4,1) 42 0 100 Second derivative(2,8,8,1) 42 0 100 6 years old
(28 types) Primitive Derivatives (0,0,1,1) 28 0 100 First derivative(1,4,4,1) 28 0 100 Second derivative(2,8,8,1) 28 0 100

In the above results, when 147 types of red ginseng were divided into 4-year, 5-year, and 6-year roots with different cultivated lotus roots, the raw spectrum, the first and second derivative spectra of each lotus root, and the first and second derivative spectra and a statistical multivariate analysis method. The result was obtained that it is possible to discriminate the lotus root of red ginseng by three-dimensional stereoscopic analysis through principal component analysis and the least-squares discriminating analysis method for deformed parts.

2) Predictive model verification

On the other hand, when the prediction model developed using the modified partial least-squares discrimination analysis method is applied to the daily analysis for the identification of lotus roots of red ginseng samples, 22 kinds of 4-year-old red ginseng root to determine whether it can actually be used as a predictive model for lotus root identification Table 11 below shows the results of verifying the accuracy when applied to unknown samples of , 25, 5-year-old, and 17, 6-year-old roots.

Math treatment SD Bias r ² SEP(C) slope RSP 0,0,1,1 0.797 0.083 0.872 0.297 0.918 2.684 1,4,4,1 0.761 0.019 0.858 0.295 0.953 2.580 2,8,8,1 0.794 0.011 0.837 0.323 0.903 2.458 Mean: average of used sample values
SD: standard deviation of mean
Bias: average difference between reference and NIRS values
r ² : coefficient of determination of cross-validation
SEP(C): the corrected standard error of prediction
RSP: SD/SEP(C), the ratio of SD of reference data to SEP(C) in the external validation set

As can be seen from the results of Table 11, the modified partial least squares discriminative analysis predictive model developed for the discrimination of red ginseng lotus root to which the first derivative is applied has a coefficient of determination (r ² ) of 0.858, which is the coefficient of determination at the time of developing the predictive model. Although it showed a rather low level of discrimination accuracy, as can be seen in Table 12 below, in the case of 4-year-old red ginseng, it was distributed in the random variable range of 3.54 to 4.64, so the discrimination accuracy was 90.9% (95.5% when the 0th derivative was applied), 5 In the case of year-old red ginseng, the discrimination accuracy was 96.0% by distribution in the random variable range of 4.58~5.59, and in the case of 6-year-old red ginseng, the discrimination accuracy was 100.0% in the random variable range of 5.50~6.64, and in the case of red ginseng lotus root discrimination, the first derivative was used. It was confirmed that red ginseng lotus root discrimination was possible at a reliable level of 95% or more (0th derivative is applied to discriminate 4-year-old red ginseng) in the least-squares discrimination analysis of the deformed part (see FIG. 18).

division correct discriminant number
(n) incorrect discriminant number
(n) accuracy
(%) 4 years old
(22 types) Primitive Derivatives (0,0,1,1) 21 One 95.5 First derivative(1,4,4,1) 20 2 90.9 Second derivative(2,8,8,1) 19 3 86.4 5 years old
(25 types) Primitive Derivatives (0,0,1,1) 24 One 96.0 First derivative(1,4,4,1) 24 One 96.0 Second derivative(2,8,8,1) 24 One 96.0 6 years old
(17 types) Primitive Derivatives (0,0,1,1) 17 0 100.0 First derivative(1,4,4,1) 17 0 100.0 Second derivative(2,8,8,1) 15 2 88.2

The present invention described above is not limited by the above-described embodiments and the accompanying drawings, and it is common in the technical field to which the present invention pertains that various substitutions, modifications and changes are possible within the scope without departing from the technical spirit of the present invention. It will be clear to those who have the knowledge of

Claims (11)

(a) measuring the near-infrared absorption spectrum for each wavelength by irradiating one or more ginseng samples with near-infrared rays;
(b) confirming the difference in the spectrum for each lotus root of the near-infrared absorption spectrum obtained for the one or more ginseng samples, and performing statistical analysis;
(c) determining the lotus root of one or more ginseng samples to be discriminated using the statistical analysis and the lotus root discrimination prediction model; The method according to claim 1, wherein the lotus root discrimination prediction model of step (c) is
(c-1) measuring a near-infrared absorption spectrum for each wavelength by irradiating a plurality of ginseng samples with near-infrared rays;
(c-2) arbitrarily dividing the plurality of ginseng samples for development of a predictive model for determining a lotus root and for verifying a predictive model for determining a lotus root;
(c-3) Obtain a raw absorption spectrum obtained by averaging the near-infrared absorption spectra obtained for the plurality of ginseng samples, its first derivative or second derivative, respectively, and check the peak difference according to wavelength in the spectrum for each lotus root and developing a predictive model for determining lotus root by performing statistical analysis;
(c-4) comparing the lotus root discrimination prediction model with a near-infrared absorption spectrum for lotus root discrimination prediction model verification, performing statistical analysis to verify the fit of the lotus root discrimination prediction model, and selecting a lotus root discrimination prediction model;
Lotus root identification method of ginseng and processed ginseng, characterized in that it is selected, including. The method according to claim 1, wherein the ginseng sample is a fresh ginseng or processed ginseng sample, and the lotus root of the fresh ginseng or processed ginseng is 4-year, 5-year, or 6-year root identification method. [4] The method of claim 3, wherein the processed ginseng is at least one selected from the group consisting of raw ginseng, dried ginseng, white ginseng, gok ginseng, jik ginseng, and red ginseng. The method of claim 1, wherein the ginseng sample is freeze-dried to a moisture content of 8% or less to prepare the sample. The method according to claim 1, wherein the near-infrared absorption spectrum of step (a) uses an absorption spectrum in a wavelength range of 400 to 2,500 nm. According to claim 1, wherein the statistical analysis in step (b) or (c) is Multiple Linear Regression (MLR), Partial Least Squares (PLS), Modified Partial Least Squares (MPLS): Modified Partial Least Squares) and Principal Component Analysis (PCA: Principle Component Analysis) method for determining lotus root of ginseng and processed ginseng, characterized in that it is performed by any one method. The lotus root of ginseng and processed ginseng according to claim 2, wherein when the plurality of ginseng samples are arbitrarily classified, the ratio of the number of ginseng samples for predictive model development or ginseng samples for predictive model verification is 1:0.2 to 1. Determination method. [Claim 3] The lotus root of ginseng and processed ginseng according to claim 2, wherein in the predictive model selection step, the accuracy of the predictive model is verified using a deformed partial least squares discriminant analysis predictive model to which the first derivative is applied in the discriminative predictive model. Determination method. 10. The method of claim 7 or 9, wherein the statistical analysis is a principal component analysis method in the lotus root discrimination prediction model,
A method for determining the lotus root of ginseng and processed ginseng, characterized in that the lotus root of the ginseng sample is discriminated according to the three-dimensional distribution position of each sample in the three-dimensional stereoscopic discrimination test. 10. The method of claim 7 or 9, wherein the statistical analysis is a modified partial least-squares discriminant analysis method in the lotus root discriminant prediction model,
In the case of a 4-year-old ginseng sample, it is distributed in a random variable range of 3.37-4.49, in the case of a 5-year-old ginseng sample, it is distributed in a random variable 4.50-5.49 range, and in the case of a 6-year-old ginseng sample, it is distributed in a random variable 5.50-6.65. A method for determining the lotus root of ginseng and processed ginseng.

Images