Tomato Maturity Detection in Complex Orchard Environment

This paper presents a model for detecting tomato maturity called EFC-YOLO11. It utilizes the C3k2-EMSDCEEM, C3k2-Faster, and C-BiFPN modules to enhance feature extraction, multi-scale feature fusion, and detection accuracy. The model also demonstrates strong adaptability to different conditions and good generalization ability. In experiments using the PF500 dataset, it shows a 2.5% improvement in mAP50 and a 3.7% improvement in mAP50-95. On the Chunxiang variety dataset, it improves mAP50 by 2.1 % and mAP50-95 by 2.4%. Additionally, in the purple passion fruit detection task, the model increases recall by 5.2%. The overall mAP50 reaches 86%.