摘要
如今,火灾问题是全世界人民都不得不面对的一个重大性灾难。随着经济快速发展,社会财富日趋增加,城市规模逐步扩大,消防工作的重要性就越来越突出。然而当前广泛使用的还是传统的依靠光感、烟感或者温感等物理传感器设备进行火灾预警检测,这种方法的信息单一导致范围有限,难以达到复杂环境下的火灾实时检测要求,因此引入YOLOv8网络模型对火灾进行检测。文章对YOLOv8算法和主要结构进行介绍,搭建实验环境,将图片进行标注工作,建立自制数据集,对数据集进行算法训练,再对训练好的模型进行预测,通过实验效果,进行分析数据,深入讨论火灾防护技术的未来发展方向。
Today,the fire problem is a major disaster that people all over the world have to face.With the rapid development of the economy,the increasing social wealth,and the gradual expansion of the city scale,the importance of fire protection work has become more and more prominent.However,at present,the traditional relying on physical sensor equipment such as light,smoke or temperature sense for fire early warning detection,this method of information single leads to limited range,difficult to meet the real-time fire detection requirements in complex environments,so the YOLOv8network model is introduced to detect fire.This paper introduces the YOLOv8algorithm and main structure,builds the experimental environment,annotates the pictures,establishes a self-made dataset,trains the algorithm on the dataset,predicts the trained model,analyzes the data through the experimental effect,and discusses the future development direction of fire protection technology in depth.
作者
焦瑜帆
赵建光
JIAO Yufan;ZHAO Jianguang(School of Electronic Information of hebei University,Hebei Zhangjiakou 075000)
出处
《长江信息通信》
2024年第2期72-74,89,共4页
Changjiang Information & Communications
基金
张家口市2023年市级科技计划财政资助项目,项目名称:基于深度学习的轨道安全智能监控预警关键技术研究,项目编号:2311010A
项目名称为张家口市2022年度基础研究专项,基于深度学习的数据采集与集成研究(2221008A)。
