Design Of A Remote Video Surveillance System Based On Mobile Car

With the development of security technology and people's attention to social security,video surveillance system has been widely used in all fields of the society.It plays an irreplaceable role in security defense,traffic safety,production management and other fields.The traditional video surveillance system often uses a fixed camera to take photos.The staff have to stare at the screen all the time to keep on track.This mode has the following limitations: firstly,multiple cameras need to be installed to extend the range of monitoring,which leads to a fairly high cost;secondly,the judgment of abnormal behavior depends on labour,which always cause low reliability and high workload.Hence,the applications of the traditional system are limited to simply monitoring,storage and playback,lacking of real-time analysis and image recognition for the scene of interest.As the result,it cannot provide real-time alarm for abnormal events.Furthermore,the traditional storage method based on database cannot satisfy the increasing requirements for high-definition long-term surveillance.Thus,there are intensive demands on improving the flexibility of monitoring mode,the capability of storing massive video data,and the efficiency of intelligent analysis for video content.In recent years,the development of intelligent terminal technology,image and video recognition technology and cloud storage technology makes the video surveillance system move towards intelligentization,big data storage,machine vision and etc.Aiming at the two disadvantages of the traditional video surveillance systems,a remote video surveillance system based on mobile car for indoor surveillance applications is designed in this project.In the system,camera is mounted on a mobile car.The car can patrol indoors follows a predefined route.The video or image data is uploaded to the cloud server by WiFi connection.Users can access the cloud server by different terminals to watch the video information in real time,or query the history data stored in the cloud.At the same time,with the benefit of online image processing technology,the system can detect moving target against dynamic background so as to find suspicious target in real time.For this purpose,a corner detection algorithm which combines the feature of line segments is proposed.The algorithm can detect the background information which is characterized by corners and line segments for indoor scenario.Then,affine principle isused to compensate the background change that caused by the movement of mobile car.Therefore the system can detect abnormal targets accurately and alarm in time.Simulation and experimental results show that the system can fulfill the remote real-time video surveillant task well.And it can detect abnormal targets in high accuracy and give an alarm in time.The system can be widely used in the indoor surveillance scenario such as unattended warehouse,office,home and other indoor environments.And it will have a broad application prospect.