AUTHOR=Liu Chunlu , Zuo Zhitian , Xu Furong , Wang Yuanzhong 

TITLE=Study of the suitable climate factors and geographical origins traceability of Panax notoginseng based on correlation analysis and spectral images combined with machine learning

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 13 - 2022

YEAR=2023

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.1009727

DOI=10.3389/fpls.2022.1009727

ISSN=1664-462X

ABSTRACT=The cultivation and sale of medicinal plants are one of the main ways to meet the increasing market demand for plant-based drugs. Panax notoginseng is a widely used Chinese herbal medicine. The environment of the cultivation area should be suitable for its growth and accumulation of bioactive components, and the phenomenon of business fraud should be avoided when purchasing. In this paper, we report the correlation between active components and climate factors based on high performance liquid chromatography (HPLC), and evaluate the influence of climate factors on the quality of P. notoginseng. In addition, the synchronous 2D-COS images of near infrared (NIR) data combined with the deep learning model were applied to distinguish the geographic origin of P. notoginseng at two different levels (district and town level). The results showed that the four main components contents of P. notoginseng are negatively correlated with the annual mean temperature and the temperature annual range. The lower annual mean temperature and the temperature annual range are favorable for the accumulation of the contents of these four main components. In addition, high annual precipitation and high humidity are suitable for the content accumulation of NG-R1, G-Rg1, and G-Rb1, while G-Rd is the opposite. For the geographic origins, the classifications at different geographic levels (different districts and towns) could be successfully distinguished through synchronous 2D-COS images combined with the ResNet model. The model accuracy of the training set, test set, and external validation is achieved at 100% and the cross-entropy loss function curves are lower. It proved the potential feasibility of the proposed method in P. notoginseng geographic traceability, even when the distance between sampling points is small. The findings of the present study could provide a reference for the cultivation of P. notoginseng in the future while ensuring the quality of P. notoginseng, and providing a reference for alleviating the phenomenon of market fraud.