Scholarlink

<p>In construction site accidents, the fatality rate from falls due to not wearing safety belts remains high. To address this issue, we propose a lightweight safety belt detection model, FMS-YOLO, based on a fused mixed local channel attention (MLCA) mechanism. FMS-YOLO integrates the FasterBlock module, MLCA mechanism, and Shape-IoU loss function into the YOLOv5s model to enhance detection accuracy and reduce computational load. Experimental results show that FMS-YOLO achieves an mAP@0.5 of 95.4% on the private dataset and 96.7% on the public dataset, with only 5.4&#xa0;M parameters and a model size of 11.2&#xa0;MB. Compared to the YOLOv5s model, FMS-YOLO improves mAP@0.5 by 3% and reduces the number of parameters by 1.2&#xa0;M. This model meets the real-time detection requirements for monitoring the safety belt usage of workers operating at heights in construction site surveillance videos.</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

FMS-YOLO: a lightweight safety belt detection algorithm for high-altitude workers based on attention mechanism and efficient architecture

  • 期刊论文

  • 发表日期：

    2025年4月1日

查看全文
  • Fangfang Lu,
  • Sangyu Yao,
  • Guxue Sun,
  • Tong Zhou,
  • Yijie Huang

摘要

In construction site accidents, the fatality rate from falls due to not wearing safety belts remains high. To address this issue, we propose a lightweight safety belt detection model, FMS-YOLO, based on a fused mixed local channel attention (MLCA) mechanism. FMS-YOLO integrates the FasterBlock module, MLCA mechanism, and Shape-IoU loss function into the YOLOv5s model to enhance detection accuracy and reduce computational load. Experimental results show that FMS-YOLO achieves an mAP@0.5 of 95.4% on the private dataset and 96.7% on the public dataset, with only 5.4 M parameters and a model size of 11.2 MB. Compared to the YOLOv5s model, FMS-YOLO improves mAP@0.5 by 3% and reduces the number of parameters by 1.2 M. This model meets the real-time detection requirements for monitoring the safety belt usage of workers operating at heights in construction site surveillance videos.