Live TV outside broadcasts that combine real action and computer-generated images could become possible for the first time, thanks to a new system being developed at Oxford University with funding from the Engineering and Physical Sciences Research Council (EPSRC).
In use, the new system is able to work out in real-time where a camera is and how it is moving, simultaneously constructing a detailed visual map of its surroundings. This enables computer graphics to be overlaid accurately onto live pictures as soon as they are produced. Previously the blending of live action and computer-generated images has only been possible in controlled studio environments.
The system itself comprises a mobile video camera connected to a laptop computer, which analyses the images it receives using software developed by the researchers. As the camera moves, the system picks out landmarks as reference points and makes a map of their 3D locations against which to measure its position.
The challenge is to estimate accurately the camera’s position and the layout of its surroundings at the same time – a task known as Simultaneous Localisation and Mapping (SLAM).
Dr. Ian Reid and Dr. Andrew Davison of Oxford University’s Department of Engineering Science are leading the project.
“This localisation and mapping technology turns a camera into a flexible, real-time position sensor. It has all kinds of potential applications,” Davison said.
As well as TV and video applications, the technology under development could provide low-cost, high-performance navigation for domestic robots. It could also be incorporated into video games or wearable computing, e.g. for use in dangerous environments, where it could confirm the wearer’s location and allow relevant guidance to be overlaid onto their view of surroundings.