Ginsenoside Rg2 content prediction in Panax ginseng based on the fusion of hyperspectral wavelengths combined with chemometric analysis

Abstract

The content of ginsenosides plays a crucial role in determining the quality of Panax ginseng C. A. Meyer. In order to clarify the potential applicability of hyperspectral imaging (HSI) for fast detecting the quality of ginseng, this study investigated the use of HSI combined with chemometric models to predict the ginsenoside Rg2 content. The results indicated that the prediction efficiency can be improved through complete wavelength fusion of different shots and the effective wavelength fusion obtained by different selection methods. Notably, the integrated utilization of the partial least squares regression (PLSR) algorithm and the fusion of effective wavelengths exhibited the best results in Rg2 content prediction, with the maximum coefficient of determination (R2) and relative percent deviation (RPD) values of 0.939 and 3.35, respectively, as well as the lowest root mean squared error (RMSE) value of 24.2 μg/g, indicating excellent model performance. This study provides valuable insights for the development and application of portable HSI equipment for rapid and non-destructive testing of Panax ginseng quality.