ATR-FTIR combined with chemometrics to distinguish geographical indications from non-geographical indications <i>Gastrodia elata</i> Bl

Gastrodia elata Bl ( G. elata ), a medicinal and edible homologous variety, has been artificially cultivated in different regions of China to meet the growing demands of human beings. In this study, attenuated total reflection/Fourier transform infrared spectroscopy (ATR/FTIR) combined with chemometrics [Principal component analysis (PCA), Partial least squares discrimination (PLS-DA), Support vector machines (SVM), and Data-driven soft independent modeling of class analogy (DD-SIMCA)] was used to differentiate between geographical indications and non-geographical indications of G. elata . PLS-DA, after the application of SNV+SD spectral preprocessing, achieved 100% accuracy on the training set and 88.89% on the test set, respectively. Under SG+SD conditions, SVM outperformed PLS-DA with 100% training set accuracy and 94.74% for the test set. A ResNet model that used synchronous 2DCOS data successfully distinguished G. elata from Yunnan and Guizhou, achieving 100% accuracy across training, test, and external validation sets. These findings support that ATR-FTIR and chemometrics can be utilized to effectively identify the geographical origin of G. elata , with potential applications for other medicinal and edible plants.