基于改进YOLOv8的火灾检测算法研究
Research on fire detection algorithm based on improved YOLOv8
焦瑜帆 1赵建光1
作者信息
- 1. 河北建筑工程学院,河北 张家口 075000
- 折叠
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.
Abstract
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 expan-sion of the city scale,the importance of fire protection work has become more and more promi-nent.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 en-vironments,so the YOLOv8 network model is introduced to detect fire.This paper introduces the YOLOv8 algorithm 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.
关键词
火灾检测/yolo/目标检测Key words
fire detection/yolo/Object detection引用本文复制引用
基金项目
张家口市2023年市级科技计划财政资助项目(2311010A)
张家口市2022年度基础研究专项(2221008A)
出版年
2024
NETL
NSTL
万方数据