Getting graphic detail

Oxford-based image-analysis expert 2d3 has signed three research deals with the Ministry of Defence to improve automated data collection from unmanned aerial vehicles (UAVs).

The deals, worth £150,000, will largely focus on improving the group’s TacitView software suite, which has been designed to exploit information from different surveillance tools to produce actionable intelligence.

Julian Morris, deputy chairman of 2d3’s parent company, OMG, said that use of such a system had become increasingly crucial in today’s military environment. ’I’ve heard an American general describe the situation for the US as swimming in sensors but drowning in data.’

He added: ’What we’re trying to do is develop a piece of software that runs on a laptop that can take the images from a UAV and provide intelligence analysts with an overview of what is occurring in real time… This will be particularly useful for the army and the air force, which record tens of thousands of videos a month in Afghanistan.’

The first research project will focus on the design of an archive system that can retrieve data on the basis of time, location and target. The company hopes that the system will be as simple to use as navigating a television freeview service and is currently in the process of improving the usability of its existing design.

A further project aims to stitch together large images from a sequence of video frames to produce an area mosaic. Morris claims that this will allow analysts to examine motion imagery without looking at each individual frame and could create possibilities for transmitting detailed information over a limited bandwidth datalink.

Alongside this, it is hoped work on reconstructing terrains from video data will improve understanding of recent changes in the layout of buildings and terrain. The technique, known as structure from motion (STF), will help to generate maps that are three dimensional by analysing the motion of an object over time.

Morris said: ’This does not have to just be used for a military application. We have also got customers looking at flooding in the UK, and others in the US that are examining forest fires. In all these scenarios, the common goal is to process imagery as fast as possible, and present it in a format that is as simple as possible. We are currently working with several UK universities and once research is complete we hope to integrate the technology into products within a year.’