|KUSUMANINGRUM, DEMI - Chungnam National University
|LEE, HOONSOO - Us Forest Service (FS)
|CHO, BYOUNG-KWAN - Chungnam National University
Submitted to: Journal of the Science of Food and Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/31/2017
Publication Date: 10/17/2017
Citation: Kusumaningrum, D., Lee, H., Kim, M.S., Cho, B. 2017. Nondestructive technique for determining the viability of soybean (Glycine max) seeds using FT-NIR spectroscopy. Journal of the Science of Food and Agriculture. 98:1734-1742. https://doi.org/10.1002/jsfa.8646.
Interpretive Summary: Various spectroscopic technologies have been used to assess quality attributes of many agricultural commodities. In this investigation, a vibrational spectroscopy technique called Fourier transform near infrared (FT-NIR) spectroscopy was investigated for evaluating viability (germination rates) of soybean seeds. The results demonstrated that using FT-NIR spectroscopy measurements with a multivariate spectral analysis method could predict germination rates of the soybean seeds under investigation with nearly 100% accuracy. The technique presented in this research is useful to seed industries by providing a nondestructive means to determine the viability of seeds.
Technical Abstract: The viability of seeds is important for determining seed quality. High quality seed will have high capability of germination, which is necessary to ensure high productivity. Hence, developing technology for the detection of seed viability is high-priority in the agriculture field. FT-NIR spectroscopy is one of the most popular devices among other vibrational spectroscopies. This study aims to use Fourier transform near infrared (FT-NIR) spectroscopy to determine the viability of soybean seeds. Viable and artificial aging seeds as nonviable soybeans were used in this research. The FT-NIR spectra of soybean seeds were collected and analyzed using a partial least-squares discriminant analysis (PLS-DA) to classify viable and nonviable soybean seeds. In addition, the variable importance in projection (VIP) method for variable selection combined with the PLS-DA was also employed to classify the seeds. The most effective wavelengths were selected by the VIP method, which selected 146 optimal variables from the full set of 1557 variables. The results demonstrated that FT-NIR spectral analysis with the PLS-DA method, whether using all variables or the VIP-selected variables, showed good performance in achieving high accuracy, close to 100%, for predicting soybean viability. Hence, use of FT-NIR techniques with chemometric analysis has the potential for rapidly and effectively measuring soybean seed viability.