Scholarlink

<p>Insulator defect detection is crucial for ensuring the safe and stable operation of power transmission and distribution networks. Traditional methods face challenges such as limited performance in detecting small targets and insufficient multi-scale feature fusion. To address these issues, this study proposes a novel detection model, RMC-YOLO. The RFCAConv module is introduced to dynamically capture long-range spatial information through receptive field attention mechanisms, significantly improving detection accuracy for small targets, such as micro-cracks and localized damage. The Multi-Scale Feature module (MSF) is designed to mitigate information loss during multi-scale feature fusion, enhancing the model’s performance consistency across objects of varying sizes. Furthermore, the Coordinate Attention-Assisted Semantic-Sensitive Feature Pyramid Network (CAA_SSFPN) strengthens the representation of critical features and improves defect localization in challenging scenarios. Experimental results on the RMC-IDE dataset demonstrate that RMC-YOLO achieves mAP50 and mAP50:95 scores of 89.6% and 66.1%, respectively, with a parameter count reduced to 2.04M. These findings underscore the efficiency and robustness of RMC-YOLO, providing a reliable solution for insulator defect detection and contributing to the safety and reliability of power transmission systems.</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

RMC-YOLO: An Object Detection Model for Insulator Defects

  • 期刊论文

  • 发表日期：

    2025年9月8日

查看全文
  • Xuerui Lan,
  • Xuyang Wang,
  • Lijun Liu

摘要

Insulator defect detection is crucial for ensuring the safe and stable operation of power transmission and distribution networks. Traditional methods face challenges such as limited performance in detecting small targets and insufficient multi-scale feature fusion. To address these issues, this study proposes a novel detection model, RMC-YOLO. The RFCAConv module is introduced to dynamically capture long-range spatial information through receptive field attention mechanisms, significantly improving detection accuracy for small targets, such as micro-cracks and localized damage. The Multi-Scale Feature module (MSF) is designed to mitigate information loss during multi-scale feature fusion, enhancing the model’s performance consistency across objects of varying sizes. Furthermore, the Coordinate Attention-Assisted Semantic-Sensitive Feature Pyramid Network (CAA_SSFPN) strengthens the representation of critical features and improves defect localization in challenging scenarios. Experimental results on the RMC-IDE dataset demonstrate that RMC-YOLO achieves mAP50 and mAP50:95 scores of 89.6% and 66.1%, respectively, with a parameter count reduced to 2.04M. These findings underscore the efficiency and robustness of RMC-YOLO, providing a reliable solution for insulator defect detection and contributing to the safety and reliability of power transmission systems.