An Elevator Passenger Abnormality Detection Method and Experimental System Based on Edge Intelligence
At present,the common server architecture based ro-ro elevator passenger anomaly detection system has the problem of communication delay and high cost.This paper proposes a method of ro-ro elevator passenger anomaly detection based on edge intelligence.This method uses YOLO v3-R network optimized with YOLO v3 inference speed as the target detection algorithm.The specific work content is divided into offline and online parts.The offline part includes hardware selection,YOLO v3-R network training and deployment,etc.The online part includes video collection,abnormal passenger detection and judgment.On this basis,a ro-ro elevator abnormal detection experimental system based on Jetson Nano hardware platform is developed.The system does not need network communication,has high reliability,low cost,and real-time reasoning speed can reach 1.96 frames per second.
国家科技期刊平台
NSTL
万方数据
