火灾智能监测系统的设计与实现  

Design and implementation of an intelligent fire monitoring system

作　　者：麦麦提艾力·图尔荪 冯俐 TURSUN Maimaitiaili;FENG Li(School of Information Engineering,Xinjiang Institute of Technology,Aksu 843000,China)

机构地区：[1]新疆理工学院信息工程学院,阿克苏843000

出　　处：《电子测试》2024年第2期43-49,共7页Electronic Test

摘　　要：鉴于现有的火灾监测系统在动态火情可视化方面的不足、准确度不高和误报漏报问题,设计并实现了高效的火灾智能监测系统。采用改进的YOLOv5s6算法模型,相较于YOLOv5s,YOLOv5s6通过改变网络结构优化算法,并结合动态火情可视化,实现更高的准确度和更快的监测速度,有效减少误报漏报。系统基于YOLOv5s的改进,主要实现静态图片、动态视频和实时视频监控监测,用户管理,数据库日志管理以及实时预警等核心功能。在监测准确性和响应速度效果较之前有显著提升,动态火情可视化的增加使得火灾监测更加直观和高效。系统的普适性较好,PyQT5构建的用户界面直观友好,MySQL实现了数据的持久化存储和高效管理,通过Pyinstaller打包,支持跨平台安装与部署。Considering the deficiencies in dynamic fire situation visualization,accuracy,and the issues of false and missed alarms in existing fire monitoring systems,an efficient intelligent fire monitoring system has been designed and implemented.This paper adopts the improved YOLOv5s6 algorithm model,which optimizes the algorithm by changing the network structure and combines it with dynamic fire situation visualization to achieve higher accuracy and faster monitoring speed,effectively reducing false and missed alarms.The system,based on the improvement of YOLOv5s,mainly implements core functions such as static image,dynamic video,and real-time video monitoring,user management,database log management,and real-time early warning.The system has significantly improved in monitoring accuracy and response speed,and the addition of dynamic fire situation visualization makes fire monitoring more intuitive and efficient.The system has good universality,with a user-friendly interface built by PyQT5,MySQL for persistent data storage and efficient management,and supports cross-platform installation and deployment through Pyinstaller.

关 键 词：火灾监测 YOLOv5算法 目标检测算法 计算机视觉 智能报警系统 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]