Design Of A Video Surveillance System Based On Frame Difference Method And Background Subtraction Method

With the rapid development of computer, communication, network technology and videosurveillance technology, video surveillance systems have been widely used in differentindustries and fields. However, in our current age, traditional network-based videosurveillance system has many obstacles in the limited storage spaces, network bandwidth andother serious defects with the huge amount of data transmitted. This paper aims to tackle theproblems of massive data size and network traffic largely increasing by using the one framedifference method and background subtraction method based video surveillance systemtechnology design and implementation. As simulation demonstrated, this network monitoringsystem is effective, stable and secure.This system is deployed in the cloud. Its development is convenient and easy to use. Itallows system architects to purchase different amount of storage space depending on thecustomer needs. This will have a perfect effect on the cost of storage devices and the networkvideo surveillance system. The cloud computing demand for clients is low, the data is safe andreliable, there is no limit to its expansion ability. It also allow the ease of data sharing fornetwork video surveillance, moreover security and real-time data retrieval are very significant.This paper explores the visual surveillance technology in the traditional network videosurveillance system versus an improved video moving target detection method for its efficientuse of bandwidth, and thereby reducing the amount of video storage. This paper analyzes themoving images of the three basic detection method, optical flow method, the frame differencemethod and background subtraction method, research on image change color space, comparethe advantages of the HSV color space for video surveillance. The system uses S3C6410microprocessor and Linux3.0.1of embedded real-time operating system for hardware andsoftware development platform, the software part S3C6410processor BootLoader (U-Boot)and the Linux kernel to do a more detailed analysis and transplantation, write andtransplantation of the DM9000driver. Then we establish the NFS network file system and USB camera head driver program of analysis and transplantation. The acquisition module useVideoForLinux2timely capture of the video data in the cloud-based application, making useof GAE in the cloud in the Python and JAVA language to achieve a cloud monitoring system.The system relies on the improved algorithm of moving target detection and attempts tointroduce other smart identification technology in the network video surveillance system andfinally achieve good results in the experiment. Video surveillance system deployed to theGoogle cloud, which greatly reduces the local storage load, which can be purchased ondemand, using Google cloud storage and cloud applications, reducing costs in the process. Thetraditional video surveillance system C/S structure is changed to a B/S architecture. Allpackets follow the http communication protocol to realize a complete cross-platformarchitecture application.