Research on Classification of Chinese Herbal Medicine Pieces Based on Fusion of YOLOv8n and HSV

The quality of Chinese herbal medicine pieces, as the core substance in TCM clinical prescriptions, directly determines the efficacy and safety of clinical treatment and serves as a pivotal control point in the standardization process of traditional Chinese medicine. This study presents a novel color grading framework combining the HSV color space model with the YOLOv8n object detection algorithm. HSV enables accurate separation of hue, saturation, and brightness components, reflecting color characteristics such as dryness, mildew, and vividness. YOLOv8n assists in locating the main herbal area, effectively removing background noise and enabling precise region-based color analysis. The proposed model was evaluated on a dataset of 3,879 images covering 201 herbal types, showing a high precision of 93.62% and recall of 96.23%. Experiments demonstrate that this HSV-YOLOv8n integrated model outperforms baseline algorithms, especially in detecting surface changes and identifying moldy regions. It achieves reliable classification even in visually complex or non-standard scenarios. The framework shows strong potential for real-time deployment in automated production environments, promoting the digital standardization of Chinese herbal medicine quality.