The Design And Implementation Of Intelligent Video Surveillance System Of A Scenic Area

With the rapid development of economy, tourism industry is booming, which means that the number of visitors in the scenic area is continuing to rise. This needs a higher level of scenic security work, and the role security technology plays in maintaining social stability is more important. In recent years, some national key scenic area has constructed its own video surveillance system. Its front end has a larger number of video capture camera; video matrix switch the video image to the control room video wall for real-time monitoring; controlling PTZ cameras timely can quickly track suspicious targets; digital video recorders archive massive videos. Such large-scale video surveillance system is playing an active role in improving the dynamic management of security and self-presentation.However, a large number of cameras and massive video data are challenge for the effective use of real time alarm monitoring and video data. The reasons are as follows. First, due to the limited number of screens, discovering alarm events by observing each video become very difficult and massive video data is more difficult to manage and view. Secondly, it’s easy to cause staff fatigue after long-term observation of video surveillance, which cannot detect security risks. In addition, traditional video surveillance system cannot be in "control" state. The system cannot play a preventive and warning role in suspicious and abnormal behavior.To solve these problems, basing on the characteristics of the scenic area, this paper proposed to construct intelligent video surveillance system in the key parts and important infrastructure, as well as to realize the "digital a scenic" strategic concept, which means to improve the information technology and management level of scenic area, to optimize the quality of public services. In order to enhance the capacity of the early warning and emergency treatment for the scenic area management, we proposed anomaly detection method based on cell motion feature and crowd density estimation algorithm based on Adaboost and histogram of oriented gradient. Meanwhile, multi-feature based abandoned objection detection automatically algorithm was given to detect the tourists abandoned objections and hazardous goods put by lawbreakers. Experiments show that the intelligent video surveillance system can integrate existing video and image information resources efficiently, sound the alarm timely, and take good use of the stored event information to efficiently retrieve the video data. Besides, the surveillance system has good expansibility and information sharing ability, and is the significant operating platform for scientifically managing the scenic area.