Low-cost image analysis for transport applications

The work described forms part of a collaborative project, funded for a period of two years by the SERC, the aim of which is to develop algorithms which will provide high quality data for a wide range of transport applications. The work carried out at Napier Polytechnic is described, which covers the development and testing of algorithms which simulate neural networks and enable vehicles to be detected and classified. The development of the TULIP system at the University of Newcastle is presented. The current work within TORG involves the use of the TULIP system for pedestrian tracking. Finally IMPACTS, a system developed at University College London, is described and its use in traffic surveillance discussed.<
>