Cameras that spot what kids are up to with spray paint before they make a mark; computers that identify a person from the way they walk; sensors that can see a weapon under a jacket. Katie Puckett reports on the future of law enforcement
It's 10.30pm on the housing estate of the future. A gang of kids loiters with intent – unaware that hundreds of tiny cameras are monitoring their every move and that, back in the control room, a computer knows who they are, where they live and that they have spray cans in their rucksacks. One raises a can – a gesture the computer recognises. Within seconds, an alarm has gone off in the housing office or police station and, within minutes, the youths have been caught, their identity cards scanned and details of their crime logged on a database.

This may sound like Robocop, but it's not as far off as you might think. Much of the technology upon which this scenario would rely is already whirring away in research labs. Cameras are being made that measure just millimetres across; computers can spot faces, unusual behaviour and weapons hidden under clothes. The scenario above will be technically possible within 10 years – and legal, if government thinking continues in its current vein. Since the 2001 World Trade Center attacks, anti-terrorism measures have been high on the Home Office agenda and increased surveillance is at the forefront. With increasing opportunities to gather, store and cross-reference vast amounts of surveillance data, privacy may soon become as important as antisocial behaviour on the list of tenants' concerns.

Estate of the art technology
Already, more and more CCTV systems are using digital cameras. These offer crystal-clear images a world away from the blurred pictures produced by old-style CCTV, which deteriorate dramatically in bad weather or darkness.

Existing non-digital systems may use infrared imaging at night, but the kaleidoscope of colours this generates often means little to operators.

Researchers at the University of Manchester Institute of Science and Technology have been working on something a bit more clever. "We've got a camera that can see through fog and mist. There's a little box you can use with an ordinary camera, which knows how fog affects the image, and then subtracts that in real-time," says Professor Nigel Allinson, who specialises in image engineering.

Scientists are getting even more excited about something called terahertz imaging, which achieves a similar effect to that staple of schoolboy fantasy, x-ray specs. "It uses the forgotten part of the spectrum to see through clothing and solid containers, so it can see whether people have concealed weapons or whether there are people hidden inside lorries," says Allinson.

Another problem with today's CCTV systems is that any system is only as good as the person watching it. CCTV footage is now commonly used as evidence in court, but its usefulness as a crime prevention tool is unclear. In August 2002, the Home Office published research into the effect of CCTV on crime levels. In 22 areas, only half had seen a fall in crime after installing CCTV, five had seen crime rise and five had seen no effect at all (in the final area, it was unclear what effect CCTV had had).

"Operator attention is one of the big issues," says Andy Rendell, associate programme manager at the Engineering & Physical Sciences Research Council, which funds research projects. "People have a bank of screens, and they can't possibly monitor them all. We need a way of getting significant events drawn to their attention."

An ever-vigilant computer with a very suspicious mind could do this. One scheme bidding for EPSRC funding aims to develop a program that can isolate changes in the physical environment – very useful for spotting vandalism on housing estates.

"If cameras are looking at the walls, and the walls begin to change, the trick is to find out whether a passing car is casting a shadow, or whether somebody's writing on them. The idea is that, in the long term, cameras will look at key places on estates and learn the difference," says Professor Stephen Gage, who specialises in innovative architecture at the Bartlett School of Architecture at University College London. "Then it can identify the perpetrator by taking a picture or tell the estate managers to get there quickly, or both."

Computers can even tell the difference between a fat person and a thin person with a bomb strapped round their waist

Even more ambitious are systems to spot unusual variations in human behaviour.

Gait analysis can deduce someone's state of mind from the way they are walking, and computers can even tell the difference between a fat person and a thin person with a bomb strapped round their waist. London Underground is hoping to run trials of a system that can scan crowds for panicked behaviour or individuals acting suspiciously – someone hanging around on the platform as train after train leaves may be a mugger, or they may be steeling themselves to jump. At UMIST, Allinson says a similar system is under development to cut crime in car parks: "If people go to one car, that's OK. But if they try several in a row, the computer will bring it to the attention of the operator."

Computers identifying suspicious behaviour is one thing, but the next leap brings in the science of identifying people by their physical characteristics, or biometrics. And if you think iris scanning exists only in the world of Minority Report, think again: a new primary school in Sunderland will use it in the canteen to access lunch money accounts.

CCTV can't get close enough to scan a suspect's iris, but facial recognition software could soon become widely used to nab persistent offenders. It could be very useful for housing officers enforcing an injunction or antisocial behaviour order. Professor Joshua Banfield is the director of Retail Loss Protection, which researches technologies to prevent theft from shops. "In three years, housing associations might well be able to have a database of people whose comings and goings need to be monitored," he says.

Price-wise, Banfield pegs such software at about £15,000, but adds that the CCTV system must be a compatible, digital one in the first place, which will push up the total cost. "Housing associations will have to make decisions about new investments within the next four or five years," Banfield warns, but he believes manufacturers will be so keen to get the technology out in the market that they may offer cheap deals for trials.

Cross-referencing data from many cameras can help in such surveillance. A technique that Allinson calls "magic dust" uses thousands of tiny cameras, no bigger than the size of a pencil rubber, which could be planted around an estate, communicating with each other over wireless links.

The computer department at the University of Essex are taking this even further. They have built an "intelligent" living environment, kitted out with tiny sensors that detect motion, temperature and movements and cameras that recognise human faces and behaviour. They are now bidding for a grant from the EPSRC to build a whole flat where robots will use information from the sensors to respond to the resident's needs.

There is a strong crime prevention angle to this, says the department's Dr Hani Hagras: "Any good thief can get past an intruder alarm, but here there are agents everywhere that can detect if someone is trying to infiltrate the system."

Dream future or Orwellian nightmare?
In 10 years, Hagras predicts, the technology could expand over a housing estate or a small village to monitor – and learn – everything that happens. And herein lies a potential minefield. As all the scientists admit, just because the technology is available doesn't mean it is desirable in the places where people live and work. Few of us relish the idea of being watched by a computer every time we leave our house, and targeting supposed dodgy characters before they've committed any crime is a very contentious issue.

Targeting supposed dodgy characters before they’ve committed any crime is a very contentious issue

"There will always be new, more advanced CCTV systems and the problem is that most don't comply with data protection and human rights legislation," says Gareth Crossman, head of policy at human rights campaign group Liberty. "The answer is to make sure people operating it have proper training." Crossman quotes research from privacy consultancy Insight Data Protection estimating that even now, nine out of 10 CCTV systems are not operated in accordance with the Data Protection Act. There may come a time when the biggest headache for housing staff is protecting their tenants from the abuse of their personal details.

Fears of a Big Brother future will only be exacerbated by the Home Office's evident enthusiasm for keeping tabs on our behaviour. Last Wednesday, it revealed that identity cards are to be piloted this autumn; meanwhile, the Office of National Statistics has been researching the feasibility of a national population register, the "Citizen Information Project".

Liberty opposes the home secretary's plans for a national ID card and the compilation of a database of citizens. Crossman also takes issue with part four of the Antisocial Behaviour Bill, which would allow police to impose curfews on under-16s in areas where antisocial behaviour is a persistent problem. This could be enforced particularly effectively using CCTV technology coupled with the date-of-birth information on ID cards.

Professor Banfield predicts that housing officers operating such surveillance systems could find themselves facing a barrage of previously unheard-of challenges. "Instead of 'what do I want to do?', it will be 'am I allowed to do it?'," he says. "Housing providers will have to become experts in data protection and human rights legislation, as retailers have had to do."

Housing providers may face a dilemma that litigation-wary retailers are already up against: whether it is better to ignore a real criminal than to wrongly accuse an innocent person.

Technology in society
Advances in surveillance techniques will not happen in a vacuum, however. The homes of the future will incorporate the latest advances in materials science, such as unbreakable glass and graffiti-resistant paint, and hopefully estates will be designed to eliminate areas vulnerable to crime.

Meanwhile, the budding graffiti artists of tomorrow will have new hobbies and entertainment opportunities open to them – which their landlords may or may not like. They will also be subject to new tenets of social policy, and it could be that the current trend towards criminalising behaviour considered unpleasant is not the answer.

Researchers at the Jill Dando Institute of Crime Science at UCL are looking into exactly this issue. "What will young people be doing in 20 or 30 years' time that doesn't annoy everyone?" says the institute's director, Professor Gloria Laycock. "What can we find that is so compelling that young people will want to get involved?"