基于改进YOLOv5的输电线路鸟巢检测方法研究

输电线路上的鸟巢会对电力设备的安全运行构成威胁，甚至影响整个电力系统的稳定性。针对复杂场景下输电线路鸟巢检测方法适用性较差的问题，提出一种基于改进YOLOv5的输电线路鸟巢检测方法。该方法结合通道注意机制和空间注意机制设计特征平衡网络，以通道权值和空间权值作为引导，实现检测网络不同层次特征之间语义信息和空间信息的平衡。同时，为了避免因网络层数增加导致特征信息不断被弱化的问题，设计特征增强模块以捕获与鸟巢相关的通道关系和位置信息。最后，利用输电线路无人机巡检图像建立鸟巢数据集进行训练和测试。实验结果表明，所提出的输电线路鸟巢检测方法具有较强的泛化能力和适用性，同时也为电力图像缺陷检测提供技术参考。

Bird nest detection method for transmission lines based on improved YOLOv5

The bird nests on transmission lines can pose a threat to the safe operation of power equipment and even affect the stability of the whole power system. To address the problem of poor applicability of transmission line bird nest detection methods in complex scenarios, an improved YOLOv5-based transmission line bird nest detection method is proposed in this paper. This method first designs a feature balancing network by combining a channel attention and spatial attention mechanism, and uses channel weights and spatial weights as a guide to achieve the balance of semantic and spatial information between features in different levels of the detection network. To avoid the continuous weakening of the feature information because of the increase of network layers, a feature enhancement module is proposed to capture the channel and location information related to the bird nest. Finally, transmission line UAV inspection images are used to build a bird nest dataset for training and testing. The experimental results show that the proposed transmission line bird nest detection method has strong generalization capability and applicability, and also provides technical reference for power image defect detection. This work is supported by the National Natural Science Foundation of China (No. 62102373, No. 61873246, No. 62072416 and No. 62006213).