近红外光谱在西洋参原材料精选上的应用.pdf

2011 年 8 月 科学仪器服务民生学术大会论文集 August 2011 近红外光谱在西洋参原材料精选上的应用 1 1 1 2 2 1 戚淑叶 黄亚伟 王灵灵 Jacqueline J SHAN Lei LING 韩东海 1、中国农业大学食品科学与营养工程学院，北京 100083 2、Afexa Life Sciences Inc. Edmonton ，T6N1G1，Canada 摘 要：本文样品均用积分球漫反射方式采集光谱。首先将原始西洋参和人参光谱采用 MWPLS 选择特征波段，建立 LS-SVM 品种鉴别模型，判别率为 100%。然后将样品粉碎，人参以 5%、10%、15%、20%掺入西洋参，于 4000-7000cm-1 建立 PCA-DA 掺杂判别模型，判别率 97.2% 。此研究有助于制药企业实现西洋参原材料实时、快速精选，保效节资。 关键词：西洋参 原材料 近红外光谱 精选 Application in Sifting of American Ginseng as Raw Material by Near Infrared Spectroscopy Qi Shuye Huang Yawei Wang Lingling Jacqueline J SHAN Ling Lei Han Donghai (1 College of Food Science and Nutritional Engineering, China Agricultural University, Beijing 100083 2 Afexa Life Sciences Inc. Edmonton, T6N1G1, Canada) Abstract: The Near Infrared spectra were optimal to use integrating sphere diffuse reflective mode. Firstly, Moving Window Partial Least-Squares regression( MWPLS) was used to pick out characteristic wavebands of original spectra of American ginseng and Panax ginseng. Then, the LS-SVM model was built in the appointed wavebands to differentiate American ginseng from Panax ginseng with a 100% accurate rate. Afterwards, samples were comminuted, and American ginseng powder were mixed with 5%, 10%, 15%, 20% Panax ginseng powder respectively. The PCA-DA model was built at 4000-7000cm-1 to distinguish the adulterant American ginseng with a 97.2% accurate rate. This research contributes to sift raw material of American ginseng timely and quickly for pharmaceutical factor