Although the names and addresses of most passengers will not be known, the software will identify where and when each relevant passenger entered the network, how they got to the incident, and an image that can be used for facial recognition and, in the case of very serious incidents, for police call-centre and casualty management.

"It determines which passengers are relevant to any incident and retrieves an image or video of those specific passengers from surveillance footage captured when they enter or exit the rail network. This will generally be a tiny fraction of the million or so passengers that would be travelling on the network at the time of a peak-hour incident," Haddy says.

"It then isolates snippets of video from the entire surveillance camera network that are highly likely to contain images of those relevant passengers throughout their individual journeys -- irrespective of where the incident itself is located. The vast majority of surveillance video is discarded as soon as the software determines that it is irrelevant to a particular incident.

It is then realistic for automatic and semi-automatic processing to be performed on the remaining surveillance video footage to prove each relevant passenger's whereabouts throughout their entire journeys."

Haddy says correlation of results generated by the tracking software has yielded some "useful 'every-day' crime fighting benefits" such as detecting pick-pocketing incidents which are considered to be low-end crime but "have a significant effect on the public's perception of a rail network's safety".