SMART SURVEILLANCE

How smart CCTV uses data to scan and analyse a scene, identifying objects and people in milliseconds

- AILSA HARVEY

Built into architecture or tucked out of sight, surveillance cameras can go unnoticed.

However, in many public spaces, such as town centres, roads, public transport hubs and private properties, they're always watching. These cameras remain in operation 24 hours a day, keeping a record of daily life to help deter illegal activity and antisocial behaviour or provide proof when it occurs.

The earliest closed circuit television (CCTV) could only livestream activity and required a security guard or police officer to observe multiple screens simultaneously in real time. The technology served as a way for one person to monitor more than one location at once, but didn't use data to make sense of the images it was streaming. Instead, it was up to human observers to notice suspicious behaviour and alert police to illegal activity.

Today, computer engineers have built systems that almost entirely take over the human supervision role. As sequences of events play out in front of the camera, a computer system can predict potential threats that might follow. It can also analyse the faces of people in crowds to help find missing people or track suspects of crime.

For collecting and processing this data, modern intelligent CCTV has a range of sensors and camera types to collect clear imagery and extract as much information as possible. This includes infrared cameras for night surveillance, which produce imagery by detecting infrared wavelengths that are invisible to the human eye. The camera then converts the infrared signature into a visible image to 'see in the dark'.