Information Technology Reference
In-Depth Information
CCTV surveillance is often justified on the basis that it can be used to prevent
terrorist attacks, but the reality is somewhat more banal. As [26] points out, those
committing acts of terrorism are usually unknown before the act takes place, and
it is very difficult to conduct surveillance of someone whose identity is unknown.
There are few examples of CCTV surveillance being used effectively for real-time
detection or prevention of crime.
It has been suggested that the effectiveness of CCTV surveillance can be im-
proved by tracking known individuals as they move through a surveillance network.
One possible application could be to follow known hooligans as they enter and move
about a football stadium. This paper will consider the design of a video surveillance
network which could track the movement of known individuals in real time.
1.1
Real-Time Video Surveillance
Let us consider a specific application area — video surveillance within a Secure
Corridor; for example, within an airport or other secure building. Airports can be
considered as a self-contained microcosm, which can be viewed in terms of the flows
through them, and the movements and behaviours that take place in their arrival,
shopping and transit areas [15].
The airport can also be viewed as a Surveillant Assemblage . The relationship
between surveillance and mobility within an airport is considered in [22]. Until re-
cently, airport surveillance did not focus on individuals - people were disassembled
into data flows to monitor movement through the airport. Individual surveillance has
come to the fore within the last ten years. Passenger profiling is conducted osten-
sibly for security, but profiling information is also of commercial use to the airport
authorities. The combination of profiling with the ability to track the movement of
individuals is potent.
A number of airports have already installed systems to register passengers as they
enter the Secure Corridor, and track when they leave. Passengers are enrolled using
their boarding card and a digital photograph is taken. When the passenger leaves the
Secure Corridor, the photograph is compared with the passenger's face by a security
officer. This generates a transaction when the passenger enters and leaves the secure
area, but cannot track their movements within it. Airports may wish to track some
individual passengers as they move through passport control, the retail shopping
area, departure lounges and exit to their gate. Our hypothetical system could achieve
this by tracking the passenger's face as they move around the secure area.
The proposed system will need to work in real time: that is, the extraction of facial
features and comparison to the stored model for each passenger must be extremely
fast. The metric used must be reasonably robust to changes in scale, illumination,
pose and occlusion. The system will need to track individuals as they move from
the field of view of one camera to another. A distributed network of cameras will
need to communicate with a back-end database and pattern-matching system. These
requirements are discussed in more detail in section 2.
 
Search WWH ::




Custom Search