Abstract
Aiming at the problems of heavy video workload and low efficiency of staff in the video monitoring system, a target detection system model combining camera and virtual electronic fence is established, and a virtual electronic fence alarm system with low equipment cost and higher safety factor is developed. The system can set any geometric area independently, recognize the human posture of pedestrians entering the area through the video information collected by the monitoring lens, detect pedestrian intrusion and give an alarm, which plays the role of shock and warning. On this basis, Open Pose algorithm is proposed to further improve the accuracy and rapidity of the system. Experimental results show that this method is feasible and effective.
Publisher
Darcy & Roy Press Co. Ltd.
Reference11 articles.
1. Xu Lanfei, Qian Xuejun. Three-dimensional simulation research of passenger movements in subway station monitoring system [J]. Railway Computer Applications, 2017，26(6):60-64．
2. Wu Mengdi. Research on the behavioral characteristics of violent and terrorist crowd monitoring in civil aviation security isolation area [J]. Scientific Chinese,2017 (24).
3. Yuexin Wu, Zhe Jia, Yue Ming, Juanjuan Sun, Liujuan Cao. Human behavior recognition based on 3D features and hidden markov models[J]. Springer London,2016,10(3): 495–502.
4. Insafutdinov E, Pishchulin L, Andres B, et al. Deepercut: A deeper, stronger, and faster multi-person pose estimation model[C]//European Conference on Computer Vision. Springer,Cham, 2016: 34-50.
5. Naji S, Jalab H A, Kareem S A. A survey on skin detection in colored images[J]. Artificial Intelligence Review, 2019, 52(2): 1041-1087.