HSD-YOLO: A Lightweight Detection Algorithm for Cotton Leaf Pests and Diseases

With the development of artificial intelligence-related technologies, the use of deep learning techniques for cotton leaf pest and disease identification not only substantially saves labor costs, but also improves efficiency. However, traditional target detection algorithms often have a large number of parameters and computation, and are unable to recognize efficiently in the complex and crowded cotton planting environment. To address this problem, this paper proposes the HSD-YOLO detection algorithm based on YOLOv8n, which firstly takes HGNetV2 as the backbone network, secondly introduces SlimNeck into the neck network to realize the model lightweighting, and finally uses DySample up-sampling to improve the detection accuracy. The experimental results prove that the improved algorithm’s detection accuracy rises by 1.42%, and the reference number and the amount of computation decreased by 18.37% and 24.69%, respectively, which makes the detection accuracy higher than YOLOv8n with the realization of model lightweighting.