Roberts, J.M. and Charnley, D.
Parallel Visual Tracking.
1st Int. Workshop on Intelligent Autonomous Vehicles
Full text not available from this repository.
The research reported here addresses the problem of tracking independently moving objects from a moving observer in real time, using corners as object tokens (Harris and Stephens, 1988). A novel algorithm is developed to solve the local correspondence/tracking problem. This algorithm relaxes the restrictive static-world assumption traditionally made, and is therefore capable of tracking independently moving objects. A parallel (Transputer) architecture is described on which to implement the algorithm. Implementation of the algorithm using T800 transputers has shown that near-linear speedups are achievable, and that real time operation is possible (half-video rate has been achieved using 30 processing elements).
Actions (login required)