An Improved Plant Disease Detection Method Based on YOLOv5

To overcome the challenges of weak generalization, high false detections, and missed targets in multi-scale plant disease detection under complex scenarios, this paper proposes an enhanced method based on YOLOv5. The method improves YOLOv5 by incorporating a multi-scale feature fusion module, an attention mechanism, and optimizing the loss function to accurately locate and classify plant disease targets. Experimental results show significant improvements over the original YOLOv5 model in terms of mAP, precision, and recall, with enhancements of 3.99%, 1.56%, and 1.27% respectively. Additionally, the proposed method outperforms other commonly used models. This study provides an efficient and accurate approach for plant disease detection, enabling real-time monitoring of plant health and precise disease control in plant factories.