This study aims to explore an intelligent building safety monitoring system based on edge com-puting to address the challenges of intelligent building safety monitoring.Firstly,the needs and challen-ges of intelligent building safety monitoring were analyzed,establishing the purpose of this study.Second-ly,a monitoring system architecture based on edge computing was designed,utilizing Yolo and SlowFast models to implement functions such as face recognition,target detection,pose recognition,and behavior detection.Subsequently,the system's performance was evaluated through experimental testing on the UCF-101 dataset.The results show that pose recognition and behavior detection based on the SlowFast model perform well in terms of accuracy,recall,and F1 score,demonstrating the system's effectiveness and re-liability.The conclusion indicates that the intelligent building safety monitoring system based on edge computing proposed in this study effectively addresses the problem of intelligent building safety monito-ring,providing reliable technical support for building safety management.In the future,algorithms and system architecture can be further optimized to enhance system performance and reliability,meeting the growing demand for building safety management.
edge computingvideo surveillanceYolo modelSlowFast model
NETL
NSTL
万方数据
