| Intelligent video surveillance in kindergartens refers to the use of video surveillance as a carrier to monitor children's daily situation in the kindergartens.It can detect abnormal problems in time and give an alarm to ensure children's safety in the kindergarten.On the one hand,with the growing maturity of computer technology,video surveillance system has penetrated into all aspects of our lives;On the other hand,with the deepening reform in the field of education,parents will pay more attention to the growth and physical and mental health of children,hoping to avoid accidents in kindergartens.However,traditional video surveillance in kindergartens can hardly play the role of early warning and alarm.Therefore,this paper takes the video surveillance system built by myself as the carrier,and under the background of the construction of intelligent kindergarten security system,devotes itself to the research of key technologies of intelligent video surveillance,so as to realize the construction of intelligent kindergarten surveillance system.In this paper,the intelligent video surveillance system for kindergartens is realized by completing the design of image-related program design at the monitoring terminal and the development of related algorithms for children's abnormal behavior detection at the PC terminal.The main work is as follows:(1)Give the overall design of the system,build the system software development environment,complete the selection of the main hardware modules of the monitoring terminal,transplant the embedded Linux system in the monitoring terminal,and add the driver related to hardware work.(2)At the monitoring terminal,the development of image-related programs is completed,and the functions of image acquisition,processing,compression,coding,OSD information overlay,MP4 file packaging and storage are realized.In order to meet the real-time requirement of wireless transmission,a video stream subpackage transmission program based on RTSP/RTP protocol is designed at the monitoring terminal to ensure the real-time performance.(3)Aiming at the situation that background may change,the background method based on Mixture Gauss Modeling is used to distinguish the foreground and background in video image sequence,Modeling with different learning rate to reduce the interference of background noise to foreground image,and successfully extract moving objects.(4)Aiming at the problems of target tracking in the process of target tracking,such as inability to initialize automatically,target occlusion and target failure,etc.A multi-target tracking algorithm based on Camshift/Kalman is proposed,which can realize the functions of automatic initialization of target tracking,target failure processing,multi-target tracking,occlusion tracking and so on.Compared with the traditional algorithm,the tracking range is reduced,and the operation speed and correctness are improved.(5)According to the scene of children's activities in the kindergarten and the possible accidents,combined with the detection,tracking and feature extraction of moving objects,the functions of cross-border detection,intrusion detection based on closed areas and speed anomaly detection are realized.Finally,the intelligent video surveillance system designed in this paper is tested in kindergartens.The experimental results show that the image is clear,smooth,stable and has high compressibility and real-time performance.At the same time,it can accurately and quickly extract and track moving objects in video images and realize abnormal behavior detection of moving objects,which embodies the characteristics of intelligent video surveillance system.In terms of algorithm performance,the average recall rate of different scenarios can reach 94.7%.In terms of detection efficiency,the average detection speed can reach 35.6 ms/frame,which meets the requirements of real-time detection and achieves the expected goal of the design of this subject.It can be applied to the construction of safety system of intelligent kindergartens. |
PDF Full Text Request