Research on Nondestructive Identification of Sorghum Varieties Based on Portable Near-infrared Technology
YANG Liping
QIU Xiwen
YANG Qinghua
ZHAN Jiaqi
TANG Hui
LI Pao
DU Guorong
Abstract：Different sorghum varieties directly determine the quality of Baijiu,so the identification of sorghum varieties is particularly important.This study used portable near-infrared(NIR)spectroscopy combined with chemometric methods to establish a sorghum varieties identification model.Spectra of japonica sorghum and glutinous sorghum in both intact and powdered states were collected at three integration times of 25,45 ms,and 65 ms.Single and combined spectral preprocessing methods were used to eliminate interference in the spectra.Unsupervised recognition methods,PCA,and three supervised pattern recognition methods,LDA,PLS-DA,and SIMCA,were used to construct identification models for different varieties of sorghum.The results indicated that spectral preprocessing methods could effectively eliminate interference.Unsupervised methods could not achieve the identification of different varieties of sorghum,while supervised pattern recognition methods could achieve the identification of sorghum varieties,where LDA＞PLS-DA＞SIMCA.Under three integration times,the results of powdered sorghum were better than those of intact sorghum.The results of the three supervised discrimination models at 65 ms were poor,the best discrimination result achieved at 45 ms integration time,and the original spectra were able to achieve 100％discrimination.The above results indicated that the combination of portable near-infrared spectrometer and pattern recognition method could achieve accurate and non-destructive identification of glutinous sorghum and japonica sorghum.
Keywords：Glutinous sorghumjaponica sorghumnear-infrared spectroscopynondestructive testingnondestructive identificationintelligent detection
Publication Date：2025-01-19
Online Publishing Date：2026-08-28(First online date of this platform, not the publication date of the document)
Pages：7( 1-7 )