ABSTRACT
Remote surveillance in the form of closed circuit television (CCTV) has been available almost since the
invention of television. It has the great advantage in providing a remote ‘eye’ to capture the scene, which
may often be in an environment hostile to a viewer. However, a viewer is still required to analyse the
data, and one may be looking for sparse events, such as an intruder breaking into a building. Images
however can now easily be captured in a digital format which in turn can be subject to algorithmic
analysis and search techniques in a computer. This leads to automation of the analysis and viewing
burden, and is often described as ‘intelligent surveillance’. The real benefit, however, often derives as
much from the automation of the decision-making task as the analysis applied to the images.
