Corn Disease Detection Based on Improved Lightweight YOLOv5

本文提出了一种基于改进型轻量化YOLOv5的玉米叶片病害检测方法，通过采用GhostNet模块替换骨干网络并结合GSConv模块优化颈部网络，显著降低了模型的参数量和计算复杂度。实验结果表明，改进后的YOLOv5s-Ghost-GSConv-seg在保持较高检测精度的同时，参数量减少37.5%，计算量降低20.8%，推理速度提升15.6%。该模型通过多尺度特征融合设计，有效平衡了检测精度与实时性，为农业智能化中的轻量化目标检测提供了高效解决方案。This paper proposes an improved lightweight YOLOv5-based method for detecting maize leaf disease detection by introducing the GhostNet module to replace the backbone network and optimizing the neck network with the GSConv module, significantly reducing model parameters and computational complexity. Experimental results demonstrate that the enhanced YOLOv5s-Ghost-GSConv-seg maintains high detection accuracy while achieving a 37.5% reduction in parameters, a 20.8% decrease in computational cost, and a 15.6% improvement in inference speed. The multi-scale feature fusion design effectively balances detection accuracy and real-time performance, providing an efficient solution for lightweight object detection in smart agriculture applications.