Humans process images with apparent ease, quickly filtering out useless information and identifying objects based on their shape and colour. In the Four-Legged League of RoboCup, the focus of most vision systems has been on using colour to recognise objects. This single-mindedness has many disadvantages, but edge detection and shape recognition are not panaceas as their computational requirements are too high. This work focuses on a technique to efficiently combine the use of colour and edge detection in order to better recognise the field landmarks and objects found in the controlled environment of the Four-Legged League of RoboCup.
Australasian Conference on Robotics and Automation (ACRA 2004). Proceedings of the 2004 Australasian Conference on Robotics & Automation (Canberra, A.C.T. 6-8 December, 2004)