1. Study on hyperspectral monitoring model of β-glucan content in oat grains.
- Author
-
Yang, Chenbo, Song, Lifang, Wang, Danli, Hao, Shuangnan, Feng, Meichen, Zhang, Meijun, Wang, Chao, Xiao, Lujie, Yang, Wude, and Song, Xiaoyan
- Subjects
BETA-glucans ,OATS ,SPECTRAL reflectance - Abstract
The β-glucan in oat grain plays an important role in preventing cardiovascular disease. Therefore, it is of great significance to realize the rapid measurement of β-glucan content in oat grain for the rational use of β-glucan. In this study, hyperspectral monitoring models of oat grain β-glucan content were constructed, and the effects of transformation algorithm, sensitive band screening algorithm, and modeling algorithm on the accuracy of the model were studied. The results showed that logarithm(Lg) and standardized normal variables(SNV) transformation can change the value of original spectral reflectance, while first-order differentiation(FD) and second-order differentiation(SD) transformation can also change the changing trend of original spectral reflectance. multivariate scattering correction(MSC), SNV, FD, and SD transformation can improve the correlation between spectral reflectance and oat grain β-glucan content to varying degrees. When using full band spectrum to construct the model, the performance of linear modeling algorithm was better than that of nonlinear modeling algorithm. Compared with uninformative variable elimination(UVE) algorithm, successive projections algorithm(SPA) had better band screening effect. Combining all the modeling results, it was found that MSC and SNV transformation algorithms had good results. Among them, the model based on SNV transformation spectrum and SPA-multiple linear regression(SPA-MLR) reached the highest accuracy, which can realize the hyperspectral monitoring of oat grain β-glucan content. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF