As the topic name suggests this thesis paper is a great stride forward for improving physical security by improving indoor security surveillance. Physical security is a multilayered phenomenon comprising of:
• Environmental Design that deters threats
• Access control which restricts access to unauthorized entities
• Intrusion Detection Systems(IDS)
• Identification and incident verification(video surveillance)
The objective of this research was to improve indoor security by expanding the intrusion detection to support the functionality of Automatic Intruder Identification and Tracking System (AIITS) through :
• Describing an architecture for the AIITS, built upon ...view middle of the document...
Challenges for video surveillance
• Recording devices which are to be used in various corners of the inspection areas are very expensive devices.
• It’s time-intensive and labour consuming task.
• The camera footage grows exponentially in volume which makes it practically impossible to search events of interest within huge volumes of data.
• Automated video analysis demands sophisticated machinery, software intelligence and complex computations and would be easily affected by factors like bad lighting, uniformed environments that has persons with same clothing, careful alignment of multi-camera systems.
• A security solution based on multi sensor data fusion (MSDF) which automates the three tasks of intrusion detection, intruder identification and locating and visual intruder tracking.
• System Design could be summarised as shown in the below figure:
The implementation video which was shown in the presentation showed an experiment done with a PTZ camera over one of the university floors. The images were captured and video indexing based on UWB events were tracked and sent to the machines over LAN. There is a RTLS and AIITS PTZ interface that shows various options directional control and coordinates positions.
Data collection included a three minute video recording...