Accurate and efficient identification of corn leaf diseases and pests is essential for safeguarding global food security and advancing precision agriculture. Although deep learning-based object detection techniques have demonstrated considerable potential, current models frequently encounter difficulties in detecting small targets, exhibit limited adaptability in complex environments, and demand substantial computational resources. To overcome these limitations, this study presents an enhanced YOLOv8-based detection framework specifically designed for the identification of corn leaf diseases and pests. We propose a novel lightweight and efficient module, termed C2f5, which restructures the original C2f module in the backbone network to more effectively retain fine-grained features of small targets while minimizing the total number of parameters. A specialized dataset containing 4,108 annotated images across seven categories of diseases and pests was developed to facilitate model training. Experimental results indicate that the improved model achieves a mean average precision (mAP@0.5) of 84.6%, surpassing the baseline YOLOv8 by 1.9%, with notable improvements in the detection performance of small-target instances. The proposed method provides a viable and efficient solution for deployment in resource-limited agricultural settings, laying the groundwork for real-time crop health monitoring systems.
Xu et al. (Fri,) studied this question.