Terahertz spectrum-based system for identifying commodity specifications of angelica dahurica

The invention provides a system for identifying the specification of an angelica dahurica commodity based on a terahertz spectrum, and belongs to the field of medicinal material identification. According to the invention, the terahertz time-domain spectroscopy technology is used for identifying the angelica dahurica of different commodity specifications for the first time, and the system established by the invention can effectively identify the angelica dahurica of various commodity specifications: angelica dahurica from Sichuan, angelica dahurica from Keemun, radix angelicae dahuricae and Bo-angelica dahurica. Compared with a PCA-random forest classification model and a PCA-SVM classification model, the average AUC, the average accuracy rate, the average precision rate, the average recall rate and the macro F1 value of the t-SNE-random forest classification model and the t-SNE-SVM classification model established by the method are obviously improved; wherein the t-SDE-random forest classification model is op