“The research team proposed a solution based on an improved YOLOv7 tiny algorithm for the automatic recognition of slice angles in Fritillaria thunbergii. This study constructed a rich dataset of Zhejiang Fritillaria images through data augmentation and optimized the original algorithm in multiple aspects. The experimental results show that the improved algorithm has significantly improved in terms of parameter count, computational complexity, recognition accuracy, and speed. Specifically, the number of parameters was reduced to 55.4% of the original algorithm, the computational cost was reduced to 59.4%, the average accuracy was improved by 12.2% when IoU was 0.5, the average absolute error of recognition angle was reduced by 4.58 °, and the average recognition time of a single image reached 8.7ms, far faster than the average human reaction time. This study not only provides a new method for automated identification of Zhejiang Fritillaria slices, but also provides useful references for the automated processing of other traditional Chinese medicines.”