Precise Identification of Tilapia Diseases and Lesion Feature Analysis Based on Improved YOLOv11

Aquaculture is vital for global food security, but high-density farming causes aquatic diseases that hinder its large-scale, high-quality development. As a major farmed fish, tilapia’s disease detection and prevention directly impact efficiency, with streptococcosis, parasitic diseases, and others being highly harmful. Traditional manual diagnosis misses early symptoms and lacks precise lesion localization, failing to meet precision aquaculture needs. While deep learning’s YOLO algorithms enable real-time aquatic disease detection, existing models have low accuracy in tilapia disease identification and lack lesion feature descriptions. To solve these issues, this study built a hierarchical lesion feature annotation system for “disease classification - lesion localization - feature description” and proposed an improved YOLOv11 algorithm with attention mechanisms. Experimental results show the optimized model (e.g., YOLOv11+FCA with 98.6% F1-score for streptococcosis) outperforms others, aiding on-site precise diagnosis, reducing blind drug use, and boosting aquaculture efficiency.