化学
线性判别分析
主成分分析
近红外光谱
校准
偏最小二乘回归
移动设备
人工智能
分光计
化学计量学
遥感
模式识别(心理学)
分析化学(期刊)
色谱法
心理学
统计
光学
计算机科学
地理
数学
神经科学
物理
操作系统
作者
Fang Wang,Tingting Hou,Shu-Lan Luo,Chunye Geng,Cunwu Chen,Dong Liu,Bangxing Han,Leilei Gao
摘要
The confusing use of Polygonati Rhizoma (PR) and Polygonati Odorati Rhizoma (POR) poses an unpredictable threat to the health of consumers. Sensitive, nondestructive, rapid, and multicomponent techniques for their detection are sought after. In this study, a low-cost, short-wavelength (898–1668 nm), and handheld near-infrared (NIR) spectrometer combined with multivariate spectral evaluation methods was used to establish calibration models for identifying PR and POR. NIR spectra were treated with a standard normal variate (SNV) before performing chemometric approaches. Then principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) were tested for calibration model development. The PCA results showed that spectral differences existed between the two herbs. However, the evaluation techniques could not separate them with the required accuracy. The PLS-DA calibration model, on the other hand, could separate the two herbs according to their spectral information with the prediction accuracy of >98.3%. Thus, it has been proven that a rapid, green, and low-cost method to support on-site and practical inspection through a handheld NIR instrument has been established to identify PR and POR and ensure the safety of the clinical medication.
科研通智能强力驱动
Strongly Powered by AbleSci AI