%0 Journal Article
%T 面向老旧社区楼道的低功耗物联网火灾预警系统设计与实现
Design and Implementation of a Low-Power Internet of Things Fire Early Warning System for Older Community Stairwells
%A 王秋蒲
%J Journal of Sensor Technology and Application
%P 593-603
%@ 2331-0243
%D 2025
%I Hans Publishing
%R 10.12677/jsta.2025.134058
%X 针对老旧居民楼火灾烟雾报警器失效的问题，本文开发了一种基于改进YOLOv10的物联网边缘计算火灾预警系统。通过设计多尺度融合模块、双模态记忆模块，结合轻量化改进策略，在树莓派5代平台实现火灾检测模型的部署。构建三级预警机制，并与社区物联网平台和119系统联动，形成“端–边–云”协同架构。实验表明，系统在三种数据集上达到86.3% mAP，误报率低于5%，推理帧率达到24.7 FPS。研究成果验证了时序感知模型与边缘计算融合的可行性，助力老旧小区消防智能化改造。
Aiming at the problem of failing fire and smoke alarms in old residential buildings, this paper develops an edge computing fire warning system based on improved YOLOv10 for the Internet of Things. By designing a multi-scale fusion module and a bimodal memory module, combined with a lightweight improvement strategy, the deployment of the fire detection model is realized on the Raspberry Pi 5 generation platform. A three-level early warning mechanism is constructed and linked with the community IoT platform and 119 system to form an “end-edge-cloud” cooperative architecture. Experiments show that the system achieves 86.3% mAP on three datasets, the false alarm rate is lower than 5%, and the inference frame rate reaches 24.7 FPS. The research results validate the feasibility of fusion of time-series sensing model and edge computing, and help the intelligent transformation of fire protection in old neighborhoods.
%K YOLOv10，
%K 物联网，
%K 智慧消防，
%K 火灾监测，
%K 边缘计算
YOLOv10
%K Internet of Things
%K Smart Fire
%K Fire Monitoring
%K Edge Computing
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=118880