Scholarlink

<p>Foreign objects such as nests, kites, balloons, and trash pose potential risks to the safe operation of transmission lines, making accurate and real-time detection essential for UAV-based inspection. To address this issue, this paper proposes a lightweight detection model, YOLOv8AL, based on the YOLOv8 framework. Specifically, a C2f-SECM module integrating spatial attention and channel enhancement is designed to improve feature extraction capability. An S-GSConv module is introduced to enhance feature representation by replacing the shuffle operation with standard convolution and ReLU activation, while an S-VoVGSCSP module is proposed to improve multi-scale feature fusion with reduced computational cost. Furthermore, a pruning strategy is applied to achieve model lightweighting and accelerate inference. Experimental results demonstrate that YOLOv8AL achieves superior performance in terms of detection accuracy, computational efficiency, and real-time capability compared with existing methods, making it suitable for practical UAV-based transmission line inspection.</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

Real-time foreign objects detection on transmission lines based on UAV images and proposed Yolov8AL

  • 期刊论文

  • 发表日期：

    2026年4月24日

查看全文
  • Yongtao Li,
  • Shanwen Qiu,
  • Yuan Zhao,
  • Ziheng Wu

摘要

Foreign objects such as nests, kites, balloons, and trash pose potential risks to the safe operation of transmission lines, making accurate and real-time detection essential for UAV-based inspection. To address this issue, this paper proposes a lightweight detection model, YOLOv8AL, based on the YOLOv8 framework. Specifically, a C2f-SECM module integrating spatial attention and channel enhancement is designed to improve feature extraction capability. An S-GSConv module is introduced to enhance feature representation by replacing the shuffle operation with standard convolution and ReLU activation, while an S-VoVGSCSP module is proposed to improve multi-scale feature fusion with reduced computational cost. Furthermore, a pruning strategy is applied to achieve model lightweighting and accelerate inference. Experimental results demonstrate that YOLOv8AL achieves superior performance in terms of detection accuracy, computational efficiency, and real-time capability compared with existing methods, making it suitable for practical UAV-based transmission line inspection.