Design And Implementation Of Video Surveillance System Based On Face Recognition In Secret Place

The new generation of surveillance system in secret place is based on Internet and wireless sensor network, and extends the application range of surveillance system to a variety of conditions. Video surveillance system is the core component in secret place surveillance system. The combinations of video surveillance system and other techniques make secret place surveillance system more intelligent and more effective. Among all these techniques,the secret place surveillance system based on face recognition is one of the most hottest research fields.Current surveillance systems in secret place have many disadvantages. For example, they usually need staffs on duty, record more information than they actually need and have single authentication type. This paper use a method of convolutional neural network to realize face recognition and apply it in the video surveillance system, promoting the appearance of surveillance system observably. Among all the methods to realize face recognition,convolutional neural network has an advantage of computing faster and working more effective. A detailed description is as follows:(1)This system detect persons in secret place and record it in logs automaticly, current video surveillance system need staffs to record person who visited the secret place manually,this system extract person information from video information,reducing the manual operation workload and raising the efficiency of system.(2)This system support videorecording when persons detected and not videorecording when nobody detected: current video surveillance system need videorecording all the time and have a large amount of redundant information. This system combine face recognition with video surveillance system, and videorecording only when persons detected and not videorecording when nobody detected, which increase the efficiency of recording greatly.(3)This system combine face recognition with other authentication methods in surveillance system, improve the efficiency of checking personal information:current video surveillance systems usually use single method to realize authentication, which is under attack easily and make checking personal information difficultly. This system use the method of face recognition to record person information, making it easier to checking information when risks occur.This paper design and implement the streaming server to get the information of camera and realize face recognition, design and implement the client to display information of video and person information. Finally, the face recognition accuracy, system functions, system performance, and other indexes are tested, the results prove that the above system has stable performance and high efficiency.