Year
2014
Abstract
Video surveillance systems are used to monitor security sensitive area in Nuclear Power Plant (NPP). Security sensitive area includes Reactor building, Fuel fabrication plant, Enrichment facility, Fuel Storage facility, Spent fuel storage facility etc. Video surveillance basically analyses video sequences to detect abnormal or unusual activities. The ultimate goal of video surveillance system is automatic generation of alarm to assist human operator online and inspection of an event offline. Like this present video surveillance technology plays an important role in nuclear security, nuclear safeguards, and ensures nuclear safety. The main objective of safeguards is the timely detection of diversion of significant quantities of nuclear material, and surveillance or monitoring with respect to safeguards is defined as instrumental observation to indicate or detect the movement of nuclear material. The inherent part of nuclear safety is physical protection of nuclear facilities and nuclear materials. Visual monitoring of nuclear facilities and nuclear materials is inherent part of Physical protection System (PPS). The main objective or goal of Physical Protection System (PPS) is to prevent unauthorized entry or access to nuclear facilities and nuclear materials. It also prevents radiological sabotage of nuclear facilities and theft of nuclear materials. For an effective Physical Protection System we need a fully automatic or smart video surveillance system. It should respond fast, reliable and should perform on robust algorithms for moving object detection, classification, tracking, and activity analysis. In this project a surveillance system is presented that supports a human operator by automatically detecting abandoned objects and drawing the operator’s attention to such events. It consists of four major parts: Moving object detection is the basic step for further analysis of video. In moving object detection moving objects are segmented from stationary background objects. The commonly used techniques for object detection used are background subtraction, statistical methods, temporal differencing and optical flow. Object classification categorizes detected object into a predefined classes such as human, vehicle, etc. It is necessary to differentiate or classify between object so that it is very easy to track and analyze action reliably. There are two main approaches toward moving object classification: shape-based and motion-based methods. The ???? next step in the video analysis is tracking, in which frame to frame progress or changes are marked of detected objects. This procedure provides identification of segmented region and generates information about the object in monitored area. The tracking step output is normally used to support and enhance motion segmentation, object classification and higher level activity analysis. The final step of the fully automatic or smart video surveillance systems is to recognize the behaviors of objects and analyze behavior to detect abnormal or unusual activities. The output of these algorithms can be used to make Physical Protection System (PPS) more effective and efficient by providing human operator information to make decision more accurate and in less time and stored data that can be used for offline inspection later.