Scholarlink

<p>Ice coating on transmission lines is a common phenomenon faced by the power system, which has a significant impact on the reliability and stability of power supply. Regarding the irregular shape, dense distribution, large quantity, and low recognition precision of ice-covered areas on transmission lines, we propose a study on icing recognition algorithm for transmission lines based on frequency domain sensing feature fusion network. Firstly, the adaptive focusing Fourier module (AFFM) is proposed, which accurately identifies and enhances the features of irregular ice cover shapes through adaptive processing in the Fourier frequency domain. Secondly, the frequency domain sensing feature enhancement module (FSFE) is designed to enhance the extraction of high-frequency boundary information from images, thereby improving the precision and adaptability of the model. Then, the bidirectional feature fusion pyramid Concat (Bifpncat) is introduced to enhance detail attention while preserving global information, thereby improving recognition precision and robustness. Finally, the identified results are combined with actual data to determine the thickness of icing, providing a new approach for calculating the icing thickness of transmission lines.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Research on icing recognition algorithm for transmission lines based on adaptive frequency domain sensing feature fusion network

  • 期刊论文

  • 发表日期：

    2025年12月3日

查看全文
  • Siyuan Zhou,
  • Chao Ji,
  • Zhenyu Li,
  • Huan Wang,
  • Yujie Zhou,
  • Ye Zhang,
  • Long Zhao,
  • Yongcan Zhu,
  • Jingyu Mu

摘要

Ice coating on transmission lines is a common phenomenon faced by the power system, which has a significant impact on the reliability and stability of power supply. Regarding the irregular shape, dense distribution, large quantity, and low recognition precision of ice-covered areas on transmission lines, we propose a study on icing recognition algorithm for transmission lines based on frequency domain sensing feature fusion network. Firstly, the adaptive focusing Fourier module (AFFM) is proposed, which accurately identifies and enhances the features of irregular ice cover shapes through adaptive processing in the Fourier frequency domain. Secondly, the frequency domain sensing feature enhancement module (FSFE) is designed to enhance the extraction of high-frequency boundary information from images, thereby improving the precision and adaptability of the model. Then, the bidirectional feature fusion pyramid Concat (Bifpncat) is introduced to enhance detail attention while preserving global information, thereby improving recognition precision and robustness. Finally, the identified results are combined with actual data to determine the thickness of icing, providing a new approach for calculating the icing thickness of transmission lines.