KR101922447B1

KR101922447B1 - Method for discriminating the cultivar of forage seeds using near-infrared spectroscopy

Info

Publication number: KR101922447B1
Application number: KR1020150158919A
Authority: KR
Inventors: 박형수; 이상훈; 이기원
Original assignee: 대한민국
Priority date: 2015-11-12
Filing date: 2015-11-12
Publication date: 2018-11-28
Also published as: KR20170055753A

Abstract

The present invention relates to a method of discriminating the breeds of Italian ragras, ferrany riaglas or tall fescue seeds, and it is possible to quickly and accurately determine the breeds of Italian raglas, perennial rice or tall fescue seeds using a near-infrared ray spectroscope. The method of the present invention is capable of nondestructive analysis of seeds itself without crushing seeds or chemical pretreatment, thereby making it possible to quickly and easily distinguish varieties of Italian ragras, perennial ricegrass or tall fescue seeds, Establish a transparent herb seed distribution system, such as preventing the distribution of bad seeds, and promote distribution revitalization.

Description

Method for Discriminating Cultivars of Pasture Seeds Using Near-Infrared Spectroscopy [

The present invention relates to a method for identifying a variety of grass seeds using a near-infrared spectroscope.

Recently, in order to stabilize the supply and demand of unstable grain feed due to intensified fluctuation of international grain prices or weather changes, and to alleviate the burden of feed costs for farms, the government has implemented various support policies have. In Korea, as the base for the production of forage crops has expanded and the cultivation area has increased sharply, the import of seeds from forage crops from abroad has also increased every year.

Choosing the best varieties and breeds for the production of good quality forage increases the productivity of the forage and eventually increases the profit of the farmers. Therefore, the importance of selecting superior varieties and breeds is increasing, and the superior varieties and breeds are sold at high prices . However, in the domestic forage seed market, seeds with poor quality such as germination defects, seeds with mixed breed and other varieties, or seeds with unspecified varieties and breed are distributed, resulting in damage caused by defective seeds I'm starting.

Some seeds for the production of forage crops, especially the seeds of grassy forage crops, are difficult to distinguish between the species and breed, because they have similar morphological characteristics, and the breeding and morphological characteristics of the seeds are distinguished . In particular, the major winter feed crops in Korea, Lolium ( Lolium Multiflorum Lam.) has similar genes in genus like L. perenne L. and it is very difficult to discriminate between species and breeds because they can hybridize with each other. In addition, although Festuca arundinacea is different from the above-mentioned Italian ragas and ferroni raglas, its morphological characteristics are similar, and it is difficult to discriminate the cultivars with the naked eye.

In order to solve the problem of discrimination of these varieties, many researchers have attempted to approach the discrimination of varieties using biochemical analysis. Studies have been carried out using PCR-based DNA markers to discriminate cultivars using specific proteins to discriminate genetic distance from seeds, to identify varieties in various economic crops and to control seed quality. In addition, SSR (Simple Sequence Repeat) technique has recently been used as a breeding discrimination technique in various crops such as wheat, barley, rice, and soybean. However, biochemical analysis for identification of breeding species and breeds requires a long analysis time for identification and is a very limited method for analyzing a large amount of samples in the field.

Recently, several studies have been reported on the quality evaluation of seeds using NIRS (Rapid Spectroscopy), which is a rapid analytical method, in the identification of varieties and country of origin. Velasco, L., C. Mollers. (2002. Nondestructive assessment of protein content in single seeds of rapeseed (Brassica napus L.) by near-infrared reflectance spectroscopy. Euphytica 123: 89-93) Seed oil, protein and fatty acid composition, and 12 varieties of wheat cultivated in Europe were identified. In addition, many research results have been reported in Korea to determine the origin of agricultural products using near infrared spectroscopy.

Pasture and forage crop seeds are small in size and morphologically similar, so it is very difficult to identify with the naked eye, and it is especially difficult to identify the seed varieties. In fact, there have been many cases of damage to feed productivity and budding system due to the difficulty in identifying seed breeds in feed production sites. Accordingly, it is required to develop a method for quickly and accurately discriminating various varieties of Italian rice grass, perennial rice seed or tall fescue seeds.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide a method for accurately discriminating varieties of Italian ragras, Pernarian ricegrass, and Tolsukushi seeds quickly and easily using a near-infrared spectroscope.

In order to accomplish the above object, the present invention provides a NIRS calibration model for discriminating Italian rice grass, perennial rice grass and tall fescue seeds.

The present inventor has sought to develop a discrimination method capable of analyzing in a timely manner while studying to identify the variety of grass seeds for rapid classification between Italian raglas, Pernarian ricegrass and tall fescue cultivars. The present inventors have found that the use of near-infrared spectroscopy can effectively discriminate Italian ryegrass, perennial ryegrass, and tall fescue seeds without being ground.

(S1) obtaining a near-infrared spectrum of a grass seed sample; (s2) mathematically preprocesses the gaps of the measured wavelengths of the near-infrared spectrum, and applies a differentiation method to the pretreated near-infrared spectrum to perform water treatment, and the water-treated near-infrared spectrum is measured by a partial least square discriminant analysis (PLS) -DA). The present invention also provides a method for discriminating the breeds of Italian ragras, ferroni ragras or tall fescue seeds.

In the present invention, the near-infrared spectrum in the step (s1) may be obtained using a near-infrared spectroscope. The near infrared spectrum used in the present invention can be obtained by scanning the visible region of the 680-1100 nm wavelength band, the NIR region of the 1100-2500 nm wavelength band, or the entire region of the 680-2500 nm wavelength band. Such a near-infrared spectrum wavelength band is easier to measure as it is without any pretreatment of the sample than other wavelength bands. In addition to not using a reagent, it does not require a preprocessing process such as quantification, mixing, heating and extraction of the sample It can be recovered quickly and without damage of the measurement sample, and can be used for other analysis.

The near-infrared spectrum can be obtained by scanning in all the methods commonly used in the art, and preferably by a reflection transmission method.

The spectra obtained in this manner are complex, and there are unidentified factors of change that change the superposition and physical properties of the extruded peaks. In order to solve this problem, it is preferable to derive the calibration equation after mathematical preprocessing of the spectrum before the regression analysis. As described above, the scattering effect of the light irradiated by the mathematical preprocessing of the spectrum is corrected first, and the outlier which deviates from a certain level is removed, thereby reducing the parameters of the regression analysis and increasing the reliability of the final calibration formula. In the present invention, it is possible to control a gap for each measurement wavelength in order to prevent the spectrum from being misinterpreted due to over-fitting of the obtained spectrum.

In the present invention, the discrimination method of the incorporation of the Italian raglas, the ferrarian rice or the tall fescue seed may divide the gaps of the measured wavelengths of the near-infrared spectrum by 0 nm, 4 nm, 8 nm or 16 nm.

In the present invention, in order to increase the accuracy of the calibration equation in the method for discriminating the breeds of Italian raglas, parsley ryegrass or tall fescue seeds, the differential method of the step (s2) may be applied with a first order or second order differentiation method. The differentiation method emphasizes the change of the absorption band by differentiating the spectra with the most popular spectrum preprocessing technique, and amplifies the spectral change and presents only the change. Therefore, it is very effective to remove the background line.

In the present invention, the water treatment in the step (s2) is performed by WXYZ (where W is a differential order, X is a nm measurement wavelength gap of spectrum, Y is a water treatment smooth 1 for softening the spectrum connection in the wavelength gap water treatment, , And Z means water-treated smooth 2 for smoothing the spectral shape), which can be varied according to the user's optimum conditions, but is preferably 0-0-1- 1, 1-4-4-1, or 2-8-8-1, and 0-0-1-1 or 1-4-4-1 is applied to the ferronial rice seeds. In case of 0-0-1-1, 1-4-4-1 or 2-8-8-1, it is the best water treatment method.

The calibration equation can be derived by performing various regression analysis after applying the mathematical preprocessing and the non-divisional water treatment as described above. Specifically, using principal component regression (principal component regression), which is the regression equation in which the error between the predicted value and the laboratory value is minimized, using the principal component extracted from the spectral data and the laboratory reference data . The regression analysis can be performed using principal component regression analysis (PCR), partial least squares method (PLS), or modified partial least squares method (MPLS) using the methematical analysis chemistry (Chemometrics) used in the near infrared ray spectroscopy, , It is preferable to determine whether the seeds are mixed or not through the Partial Least Square Discriminant Analysis (PLS-DA) in the method of discriminating the seeds of the perennial ryegrass or the tall fescue seed.

(S1) obtaining a near-infrared spectrum of a grass seed sample; (s2) mathematically preprocesses the gaps of the measured wavelengths of the near-infrared spectrum, and applies a differentiation method to the pretreated near-infrared spectrum to perform water treatment, and regulates the water-treated NIR spectra by partial least squares (PLS) Thereby deriving a calibration equation; (s3) mutually verifying the calibration equation; And (s4) a step of discriminating the breeds of Italian raglas seed, perennial raglas seed or tall fescue seeds using the calibration formula, and a method of discriminating the breeds of Italian ragras seed, perennial rice seed or tall fescue seeds to provide.

In the determining method, in the step (s3), the calibration equation may be mutually verified through independent verification.

In the present invention, the varieties of Italian ragas are selected from the group consisting of Greenfarm (GF), Kowinearly (KE), Kogreen (KG), Kowinmaster (KM), Kospeed (KS) and Hwasan 104 Or more can be used. Also, at least one selected from the group consisting of Accent, Bison, Tetrellite, Top Gun and the like may be used as the varieties of the Pernarian Ryegrass. The tall fescue may be selected from the group consisting of Fawn, Greenmaster (GM), Kentucky 31 (KY 31), Purumi, and the like.

The present invention can quickly and accurately determine the variety of Italian lyra grass seed, perennial rice seed or tall fescue seed using a near infrared ray spectroscope.

The method of the present invention is capable of nondestructive analysis of seeds itself without crushing seeds or chemical pretreatment, thereby making it possible to quickly and easily distinguish varieties of Italian ragras, perennial ricegrass or tall fescue seeds, Establish a transparent herb seed distribution system, such as preventing the distribution of bad seeds, and promote distribution revitalization.

Fig. 1 is a photograph showing an Italian ragras seed, a perennial rice seed, and a tall fescue seed.
2 is a photograph showing the seeds used in the method of measuring a large seed of the present invention.
Fig. 3 is the seed average NIR spectrum of the Italian lyglas varieties.
Fig. 4 is the seed average NIR spectrum of the rice varieties.
FIG. 5 is a seed average NIR spectrum for each tall fescue cultivar.
FIG. 6 is a diagram showing the results of discriminating the breeds of Italian rice species using the PLS-DA technique. FIG.
FIG. 7 is a diagram showing the results of discrimination of the varieties of the Pernarian Ryegrass seeds using the PLS-DA technique. FIG.
FIG. 8 is a diagram showing the results of discriminating the variety of the tall fescue seeds using the PLS-DA technique.

Hereinafter, embodiments of the present invention will be described in detail to facilitate understanding of the present invention. However, the embodiments according to the present invention can be modified into various other forms, and the scope of the present invention should not be construed as being limited to the following embodiments. Embodiments of the invention are provided to more fully describe the present invention to those skilled in the art.

Sample preparation of grass seed

100 of each of the six varieties of Italian Ryegrass (Greenfarm (GF), Kowinearly (KE), Kogreen (KG), Kowinmaster (KM), Kospeed (KS), and Hwasan 104 (Faent, Greenmaster (GM), Kentucky 31 (KY 31), and Purumi) were obtained from domestic seed companies.

Near-infrared spectroscopy Analysis method

SpectraStar 2500 scanning monochromator (Unity Scientific, Brookfield, USA) equipped with a rotating module was used for the near-infrared spectroscopy. The sample was placed on circular cells of polarized aluminum with quartz glass (35 mm diameter; 10 mm depth). The samples were scanned 42 times to derive the average spectra in a reflection fashion ranging from 680 to 2500 nm at 1 nm intervals. All spectra are recorded as log (1 / R ), and R is the relative reflectance. Spectra were recorded with U-CAL software version 2.04 (Unity Scientific, Brookfield, USA).

Experimental Example One: PLS -DDA method

We used the PLS-DA technique to determine the breeds of Italian ragas, perennial rice, and tall fescue, respectively. FIG. 3 shows the seed average NIR spectra of the varieties of Italian Ryegrass, and FIG. 6 and Table 1 show the results of the type discrimination. FIG. 4 shows the seed average NIR spectra of the varieties of the Pernarian Ryegrass, and FIG. 7 and Table 2 show the results of the type discrimination. FIG. 5 shows the seed average NIR spectra of the tall fescue cultivars, and FIG. 8 and Table 3 show the results of the cultivar discrimination.

For the measurement, a small ring cup with a diameter of 35 mm and a height of 10 mm was used, and a wavelength range of 680-1100 nm, 1100-2500 nm, or 680-2500 nm, 0-0-1-1 (no treatment) -4-1 and 2-8-8-1 were used.

The results of the discrimination are listed in one of three categories: 'Miss', 'Uncertain' and 'Hits'. When both limits are set at 2.5 on the input screen ('T critical limit' and 'Uncertainty factor'), then (2.0 + 'T critical limit' × SECV) or less than (1.0- 'T critical limit' × SECV) The sample with the expected value of 'Miss' was classified. (1.5 ± 'Uncertainty factor' × SECV / 2) are listed as 'Uncertain'. The remaining samples were considered as 'Hits'.

kind Wavelength range
(nm) Water treatment Identification of IRG cultivars (n = 100) accuracy
(%) Exact identification
(Hits) Uncertain discrimination
(Uncertain) Unsure
(Miss) GF
(Greenfarm) 680-1100 2,8,8,1 99 One 0 99 1100-2500 0,0,1,1 94 6 0 96 680-2500 2,8,8,1 100 0 0 100 KE
(Kowinearly) 680-1100 2,8,8,1 98 0 0 98 1100-2500 2,8,8,1 98 0 0 98 680-2500 2,8,8,1 99 One 0 99 KG
(Kogreen) 680-1100 1,4,4,1 96 4 0 96 1100-2500 1,4,4,1 98 2 0 98 680-2500 2,8,8,1 99 One 0 99 KM
(Kowinmaster) 680-1100 1,4,4,1 100 0 0 100 1100-2500 1,4,4,1 100 0 0 100 680-2500 1,4,4,1 100 0 0 100 KS
(Kospeed) 680-1100 1,4,4,1 100 0 0 100 1100-2500 2,8,8,1 100 0 0 100 680-2500 1,4,4,1 100 0 0 100 HS104
(Hwasan 104) 680-1100 1,4,4,1 98 2 0 98 1100-2500 2,8,8,1 99 One 0 99 680-2500 1,4,4,1 100 0 0 100

kind Wavelength range
(nm) Water treatment PRG cultivar discrimination (n = 50) accuracy
(%) Exact identification
(Hits) Uncertain discrimination
(Uncertain) Unsure
(Miss) Accent 680-1100 0,0,1,1 49 One 0 98 1100-2500 0,0,1,1 50 0 0 100 680-2500 0,0,1,1 50 0 0 100 Bison 680-1100 0,0,1,1 49 One 0 98 1100-2500 0,0,1,1 50 0 0 100 680-2500 0,0,1,1 50 0 0 100 Tetrellite 680-1100 1,4,4,1 50 0 0 100 1100-2500 0,0,1,1 50 0 0 100 680-2500 0,0,1,1 50 0 0 100 Topgun 680-1100 0,0,1,1 50 0 0 100 1100-2500 0,0,1,1 50 0 0 100 680-2500 0,0,1,1 50 0 0 100

kind Wavelength range
(nm) Water treatment Tall fescue variety discrimination (n = 100) accuracy
(%) Exact identification
(Hits) Uncertain discrimination
(Uncertain) Unsure
(Miss) Fawn 680-1100 1,4,4,1 98 2 0 98 1100-2500 0,0,1,1 100 0 0 100 680-2500 1,4,4,1 100 0 0 100 GM
(Greenmaster) 680-1100 2,8,8,1 93 7 0 93 1100-2500 2,8,8,1 100 0 0 100 680-2500 2,8,8,1 100 0 0 100 KY 31
(Kentucky 31) 680-1100 1,4,4,1 100 0 0 100 1100-2500 2,8,8,1 100 0 0 100 680-2500 1,4,4,1 100 0 0 100 Purumi 680-1100 1,4,4,1 92 8 0 92 1100-2500 2,8,8,1 100 0 0 100 680-2500 1,4,4,1 100 0 0 100

From the results shown in Tables 1 to 3, the method of discriminating Italian raglas, perennial rice or tall fescue seeds using the PLS-DA technique can be applied to all kinds of varieties in a wide range of wavelengths and water treatment conditions with high accuracy And confirmed that it can be discriminated.

Experimental Example 2: PLS Identification method of varieties using technique

Using the PLS technique, an experiment was conducted to determine the breeds of Italian ragras, perennial rice or tall fescue seeds. The NIR wavelength ranged from 680 to 2500 nm and water treatment was performed using 0-0-1-1, 1-4-4-1, or 2-8-8-1. Table 4 shows the result of discriminating the Italian rice species according to the wavelength range and the water treatment method.

Setting of loading value: GF = 10, KE = 20, KG = 30, KM = 40, KS = 50, HS104 = 60

Water treatment sample water Calibration equation
(Calibration) Mutual verification
(Cross validation) SEC R ² SECV R ² _CV 0,0,1,1 564 2.49 0.98 2.81 0.97 1,4,4,1 592 2.42 0.98 3.50 0.96 2,8,8,1 594 3.17 0.97 4.37 0.93

Table 5 shows the results of discrimination of the varieties of perennial rice according to the wavelength range and the water treatment method.

Set loading values: Accent = 10, Bison = 20, Tetrellite = 30, Topgun = 40

Water treatment sample water Calibration equation
(Calibration) Mutual verification
(Cross validation) SEC R ² SECV R ² _CV 0,0,1,1 198 0.65 0.99 1.02 0.99 1,4,4,1 197 0.87 0.99 1.21 0.98 2,8,8,1 199 0.96 0.99 1.43 0.98

Table 6 shows the result of discriminating the tall fescue according to the wavelength range and the water treatment method.

Setting the loading value: Fawn = 10, GM = 20, KY 31 = 30, Purumi = 40

Water treatment sample water Calibration equation
(Calibration) Mutual verification
(Cross validation) SEC R ² SECV R ² _CV 0,0,1,1 492 1.20 0.99 1.71 0.98 1,4,4,1 499 1.15 0.99 1.91 0.97 2,8,8,1 499 1.11 0.99 1.87 0.97

From the results shown in Tables 4 to 6, it can be seen that the method of discriminating the varieties of Italian raglas, Pernellagrass or tall fescue seeds using the PLS technique can discriminate each variety with high accuracy under all water treatment conditions Respectively.

Claims

(s1) obtaining a near-infrared spectrum of the grass-roots sample;
(s2) mathematically preprocesses the gaps of the measured wavelengths of the near-infrared spectrum, and applies a differentiation method to the pretreated near-infrared spectrum to perform water treatment, and the water-treated near-infrared spectrum is measured by a partial least square discriminant analysis (PLS) -DA), < / RTI >
The water treatment is WXYZ (where W is a differential order, X is the nm measurement wavelength gap of the spectrum, Y is the water treatment smooth 1 for smoothing the connection of spectra in the wavelength gap water treatment, and Z is the smoother , And 1-4-4-1 or 2-8-8-1 are applied for Italian raglass seeds and 0-0-1-1 for peranial rice seeds -1 or 1-4-4-1. In the case of tall fescue seeds, Italian raglass seeds applying 0-0-1-1, 1-4-4-1 or 2-8-8-1, How to determine the breed of the Niallayra grass seed or the tall fescue seed.

delete

(s1) obtaining a near-infrared spectrum of the grass seed sample;
(s2) mathematically preprocesses the gaps of the measured wavelengths of the near-infrared spectrum, and applies a differentiation method to the pretreated near-infrared spectrum to perform water treatment, and regulates the water-treated NIR spectra by partial least squares (PLS) , Which leads to a calibration equation,
The water treatment is WXYZ (where W is a differential order, X is the nm measurement wavelength gap of the spectrum, Y is the water treatment smooth 1 for smoothing the connection of spectra in the wavelength gap water treatment, and Z is the smoother Quot; smooth < / RTI > smooth 2 "), applying 0-0-1-1;
(s3) mutually verifying the calibration equation; And
(s4) A method for distinguishing an Italian ragass seed, a perennial rice seed or a tall fescue seed, comprising the step of discriminating an Italian ragass seed, a ferronial raglas seed or a tall fescue seed using the above calibration formula.

delete