The Study And Implementation Of Intrusion Detection Algorithm In Sensitive Area

The modern society is dense and highly complex, faced with the highly increasing unexpected events and abnormal events, consequently the difficulty and importance of monitoring is becoming more and more prominent. Especially in some sensitive security areas, such as bank, chemical plant, prisons, military bases, we need intrusion detection for the requirements of security and management in those areas, when it is discovered that some suspicious persons or vehicles breaking into the regions, the monitor system can automatically detect the invasion target, and mark the invasion trajectories, and give the alarm notification managers to process.This paper is focused on the research of intrusion detection algorithm in video surveillance system, including the moving target detection and target tracking technology. For moving target detection, the background reconstruction method is used to accurately locate motion target, considering the characteristics of the background difference method and frame difference method. In order to track the moving object, the color-based Mean Shift algorithm is used, combined with Kalman filter joined the position prediction, ensuring the tracking stability and robustness.In this paper, the functional modules of cross-border detection, forbidden zone break and climbing detection have been realized. The experimental results show that, the algorithm of intrusion detection is strongly robust, highly accurate and real-time.