Scholarlink

This article presents an intelligent fire monitoring system designed using the ROS framework and YOLO object detection algorithm. By combining ROS’s network coordination with YOLO’s high-speed visual analysis, the system facilitates fast fire detection and seamless collaboration across monitoring nodes. It comprises video capture, object detection, and alarm notification modules, with the ESP32-CAM microcontroller and OV2640 camera managing real-time video capture while YOLO performs detection. The system demonstrates strong fire detection performance with high accuracy, recall, and mAP, processing 45 frames per second for timely detection and response. However, limitations exist in dataset size and diversity. Future work could focus on expanding the dataset and exploring deeper network architectures, attention mechanisms, and adaptive learning rates to enhance performance.

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

Implementation of a Fire Detection System Based on YOLOv5

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yuh-Chung Lin,
  • Shun Nian Luo,
  • Ta-Wen Kuan,
  • Xiaodong Yu,
  • Yezhe Shi,
  • Diakonova Vasilisa

摘要

This article presents an intelligent fire monitoring system designed using the ROS framework and YOLO object detection algorithm. By combining ROS’s network coordination with YOLO’s high-speed visual analysis, the system facilitates fast fire detection and seamless collaboration across monitoring nodes. It comprises video capture, object detection, and alarm notification modules, with the ESP32-CAM microcontroller and OV2640 camera managing real-time video capture while YOLO performs detection. The system demonstrates strong fire detection performance with high accuracy, recall, and mAP, processing 45 frames per second for timely detection and response. However, limitations exist in dataset size and diversity. Future work could focus on expanding the dataset and exploring deeper network architectures, attention mechanisms, and adaptive learning rates to enhance performance.