Share Email Print

Proceedings Paper

A Real-Time Multi-Processor System For Knowledge-Based Target-Tracking
Author(s): P. D.S. Irwin; S. A. Farson; A. J. Wilkinson
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

This paper describes a multi-processor implementation of a knowledge-based system which has been developed with the aim of achieving real-time image interpretation within a specific application area, namely tracking targets in I.R. imagery. Image segmentation and feature extraction algorithms generate a reduced resolution representation of the image in terms of various features. This feature data is then fed to a knowledge-based system for image interpretation. A custom chip-set for implementation of the 'front-end' segmentation and feature extraction algorithms is described which uses a bit-level systolic array architecture to meet the high computational requirements in real-time. Reduced resolution feature arrays which are generated by this hardware are simultaneously mapped into the memory of an array of INMOS transputers. A multi-processor 'farm' architecture is described capable of implementing the first level of image interpretation algorithms in real-time. Efficient task scheduling within the multi-processor system is critical for real-time operation and an optimal algorithm for scheduling the various tasks involved in implementing the rule set, which forms the core of the knowledge-based system, is presented.

Paper Details

Date Published: 6 December 1989
PDF: 14 pages
Proc. SPIE 1154, Real-Time Signal Processing XII, (6 December 1989);
Show Author Affiliations
P. D.S. Irwin, Queen's University Belfast and Short Brothers PLC (United States)
S. A. Farson, Short Brothers PLC (United States)
A. J. Wilkinson, Queen's University Belfast (United States)

Published in SPIE Proceedings Vol. 1154:
Real-Time Signal Processing XII
J. P. Letellier, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?