By Steve Parker, PhD, and Fan Wu, PhD, RFEL
Sight is the most advanced and dominant human sense, and it is estimated that 25% of the brain is dedicated to the processing of images. The brain even prefers to ‘visualise’ problems and see information pictorially, as vision is so central to the way in which the world is perceived. It is therefore not surprising that visual surveillance is so dominant for implementing defence and security solutions. However, as with the brain, the manipulation and analysis of images comes at a heavy price, with real-time processing requiring considerable computing resources and potentially a prohibitively large power budget. This article explores two key aspects of surveillance: the digital image processing that can extend the capabilities of visual surveillance; and the processing hardware that can be used to facilitate real-time operation with an acceptable size, weight and power, plus cost (SWAP+C).