Effect of surveillance cameras on crime prevention

By Paul Wanaye Wamimbi

Basing on the recent criminal activities in Kampala capital city and its suburbs the Government is installing camera systems, and many are excepting good results, but independent research on their effectiveness is scarce.

The presence of cameras is effective in reducing crime for some, though not all, areas. The key isn't in just having cameras, but in how they are used — how many cameras are employed and where they're set up, how well they're monitored, and how well officials balance privacy concerns with utility.

In this age of hi-tech, it may be our expectations of CCTV technology are too high. People always expect that CCTV is going to do more than it's actually designed for. We tend to look at CCTV as a detection tool, but there's no real detection going on. In reality, today's CCTV systems still rely on their human operators. And humans are… only human.

The human side of CCTV, is it possible to improve operator surveillance by recognising and playing to the limitations of human biology? Active surveillance is an intellectually demanding task. It requires you to sit behind a screen and focus intensely. Really, 30 to 40 minutes is the maximum you can monitor effectively what's going on in a very complex environment. For dedicated surveillance, operators can really only focus on between 1 to 2 monitors at a time. A good operator must also have the ability to differentiate between colours, pick up on motion and speed, recognise patterns and anomalies, and have a good memory.

They must also have good observation skills, including the ability to scan, focus, and make use of their peripheral vision, and have access to good intelligence information. A good operator who understands what's going on in the environment can detect individuals who are preparing to commit crimes, intervene and prevent the crime from being committed.

You might have an operator in a control room with a couple of hundred cameras. They could have up to 20 or more screens to monitor. They could also be manning a phone and other things. In this case they are not necessarily going to see the subtle behaviours that will alert them to an incident starting to emerge."

That's where the development of new technologies such as 'intelligent CCTV' may help in the future. There's a lot of work to try and develop algorithms that, say, can detect when a person is moving in the opposite direction to the way everyone else is moving. This could be particularly useful in airports to detect if people are trying to move back into the aircraft."

But at the moment these new technologies are still a long way from being intelligent. It's hard for a computer to pick out individual people and to translate the movement of blocks of pixels into real people. And in an environment like an airport terminal where there's lots of visual noise, lots of people moving about and people sitting still for long periods of time, it's almost an impossible task."

Biometric technology such as face recognition may be another surveillance option. Face recognition is out there, but it's still very easy to defeat. Firstly, you have to have an extensive database and a controlled environment where people are face on to the camera. And your attacker could just put sunglasses on or say, in the case of a bikie, trim their beard off and thus destroy the ability of these systems to recognise a face.

So for the moment it seems we shall have to keep depending on the skill of our operators. I think that that will be the case for the next 20 years.

The writer a youth minister at the InzuYamasaaba