基于改进YOLOv5的输电线路鸟巢检测方法研究
DOI:
CSTR:
作者:
作者单位:

1.郑州轻工业大学电气信息工程学院;2.国网河南省电力公司电力科学研究院

作者简介:

通讯作者:

中图分类号:

TM726.3

基金项目:

国家自然科学基金(No.62102373, 61873246, 62072416, 62006213). 河南省科技攻关计划项目(No. 212102310053). 河南省高校科技创新人才项目(No.21HASTIT028)


Bird"s nest detection method for transmission lines based on improved YOLOv5
Author:
Affiliation:

1.College of Electrical and Information Engineering,Zhengzhou University of Light Industry;2.State Grid Henan Electric Power Reserch Institute

Fund Project:

The National Natural Science Foundation of China (General Program, Key Program, Major Research Plan)

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    输电线路铁塔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。针对输电线路鸟巢检测现有的方法适用性较差的问题,本文提出了一种基于注意力引导特征融合的输电线路鸟巢检测方法(AGFF-YOLOv5)。该方法在YOLOv5目标检测框架的基础上,通过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而丢失特征信息,提出特征增强模块以捕获与目标相关的通道信息和位置信息。最后,利用输电线路无人机巡检图像建立鸟巢数据集并进行对比实验。实验结果表明,AGFF-YOLOv5输电线路鸟巢检测方法具有较强的泛化能力和适用性,同时也为其他电力故障检测提供技术参考。

    Abstract:

    Bird’s nests on transmission line pylons can pose a threat to the safe operation of power equipment and even affect the stability of the entire power system. To address the problem of poor applicability of existing methods for transmission line bird nest detection, this paper proposes an attention-guided feature fusion-based transmission line bird’s nest detection method (AGFF-YOLOv5). Based on the YOLOv5 object detection model, the method designs feature balance network by combining channel attention mechanism and spatial attention mechanism, and uses channel weights and spatial weights as guides to achieve the balance of semantic and spatial information between features at different levels of the detection network. Meanwhile, in order to avoid losing feature information due to the increase of network layers, a feature enhancement module is proposed to capture the channel information and location information related to the target. Finally, the bird"s nest dataset is built and compared with the transmission line UAV inspection images for experiments. The experimental results show that the AGFF-YOLOv5 transmission line bird nest detection method has strong generalization capability and applicability, and also provides technical references for other power fault detection.

    参考文献
    相似文献
    引证文献
引用本文
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2022-03-27
  • 最后修改日期:2022-06-20
  • 录用日期:2022-07-01
  • 在线发布日期:
  • 出版日期:
文章二维码
关闭
关闭