AUTHOR=Zhang Silu , Wang Jingzhe , Yang Kai , Guan Minglei 

TITLE=YOLO-ACT: an adaptive cross-layer integration method for apple leaf disease detection

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 15 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2024.1451078

DOI=10.3389/fpls.2024.1451078

ISSN=1664-462X

ABSTRACT=Apple is an important economic product in China, and the yield of apples is a primary concern for fruit growers. Apple leaf diseases are one of the main factors affecting apple growth and yield. To enhance disease detection efficiency while reducing false detection caused by complex backgrounds, lighting conditions, shooting angles, or intrinsic characteristics of the diseases, this paper proposes an Adaptive Cross-layer Integration Method for Apple Leaf Disease Detection .Based on YOLOv8s, three improved modules are designed to enhance the accuracy of apple leaf disease detection, effectively mitigating the impact of external environmental factors. This method also addresses the negative issues arising from significant feature differences or similar disease characteristics, thereby enhancing the detection performance of the model. The results demonstrate that the proposed method achieves a mean Average Precision (mAP) of 85.1% for apple leaf disease detection, outperforming YOLOv5s, YOLOv8s, and other classical algorithms.The mAP improved by 2.8% compared to the baseline and exceeded the latest state-of-the-art model YOLOv10 by 2.2%. This approach effectively reduces both missed and false detections, significantly enhancing the detection and localization of diseases. It provides a new theoretical basis and research direction for apple disease detection.