Target detection algorithm for substation equipment based on YOLOv11

This paper presents an enhanced YOLOv11 algorithm specifically tailored for optimizing the detection of substation equipment targets. By incorporating the Coordinate Attention (CA) mechanism into the backbone network, our approach significantly improves the modeling of spatial and channel dependencies within feature maps. This not only boosts the model's analytical capabilities but also enhances its information fusion capacity, leading to marked improvements in detection accuracy and robustness. Additionally, we have refined the pooling layer design to strengthen the model's proficiency in handling multi-scale objects, thereby elevating object detection performance. Experimental evaluations using a substation equipment dataset demonstrate that, relative to the original version, the improved YOLOv11 achieves a 2.6 percentage point increase in accuracy, a 2.5 percentage point rise in recall rate, and a 1.4 percentage point improvement in mean Average Precision at IoU threshold of 0.5 (mAP50).