福建电脑2025，Vol.41Issue(11)：66-74,9.DOI:10.16707/j.cnki.fjpc.2025.11.013
融合YOLOv10与规则引擎的火灾隐患检测系统
Fire Hazard Detection System Integrating YOLOv10 and Rule Engine
摘要
Abstract
To enhance the intelligent monitoring capability of fire hazards,this paper constructs a multi-objective recognition system based on YOLOv10.By establishing an image dataset containing 7 types of hidden danger targets such as water bottles,cigarette butts,and electric scooters,combined with data augmentation and rule engine technology,real-time detection and semantic level warning of hidden dangers can be achieved.Testing in the Alibaba Cloud GPU environment shows that the system mAP50 reaches 0.808,mAP50-95 is 0.639,inference speed is 7.63 FPS,and some category detection accuracy exceeds 0.98.The experimental results have verified the effectiveness and real-time performance of the system in identifying multiple types of fire hazards.关键词
安全工程/深度学习/目标检测/火灾预警/智能监控Key words
Safety Engineering/Deep Learning/Object Detection/Fire Warning/Intelligent Monitoring分类
计算机与自动化引用本文复制引用
田海稳,刘原佑,桂雪峰,汪华珍..融合YOLOv10与规则引擎的火灾隐患检测系统[J].福建电脑,2025,41(11):66-74,9.基金项目
本文得到2024年黔南民族师范学院"大学生创新创业训练计划"(No.2024106700719)资助. （No.2024106700719）
