Intelligent CCTV surveillance: Advances and limitations
S.A. Velastin
Digital Imaging Research
Centre, School of Computing & Information Systems, Kingston University,
Kingston-upon-Thames, United Kingdom
The installation of
closed circuit television (CCTV) cameras in urban environments is now
commonplace and well-known. The UK leads the world with an estimated 4
million public cameras installed (20% of the world's deployment). Public
attitudes to these systems reflect the balance needed between two conflicting
requirements:
a) Concerns over invasion of privacy and fears of authoritarian control
of the population.
b) Welcoming the increased safety in public spaces and reductions in crime
and antisocial behavior.
Recent events and what appears to be the effectiveness of the CCTV infrastructure
in assisting law enforcers to understand how the events took place and
the people involved, seem to have tipped the balance towards (b), at least
momentarily.
What we should not forget however, is that such events are thankfully
very rare but that there is a cumulative significant effect of mundane
daily events that we need to deal with. For example, it has been estimated
that what in the UK is called "antisocial behavior" costs the
country around € 5000 million a year. A single London Borough (municipality)
spends annually around € 1 million to remove graffiti, 44% of women
feel unsafe at bus stops at night, a single bus company in a major city
is known to have replaced 8,000 windows in one year and a study showed
that in a single day in the UK there were around 66,000 reports of nuisance
or loutish behavior. At the same time, there are reports that when CCTV
has been installed there has been a 35% reduction in crime over 5 years.
Indeed, crime in general is decreasing while uncivil behavior is on the
increase.
It turns out that one of the major limitations of conventional CCTV systems
is the impracticality of deploying sufficient number of people to be in
front of television screens observing largely uneventful video. As long
as this is the case, CCTV will tend to remain a reactive tool. The inability
of being truly pro-active, producing timely alarms and eventually being
able to prevent incidents, is what ultimately limits these systems. As
a preamble to the main associated symposium, this talk will illustrate
some of the efforts that the research and industrial communities are making
towards realizing automated means of detecting video events involving
human activity. It will show the kind of progress made but also the current
limitations of this technology.
Paper presented
at Measuring Behavior 2005
, 5th International Conference on Methods and Techniques
in Behavioral Research, 30 August - 2 September 2005, Wageningen, The
Netherlands.
© 2005 Noldus
Information Technology bv
|