Share Email Print

Proceedings Paper

Integration Of Parallel Image Processing With Symbolic And Neural Computations For Imagery Exploitation
Author(s): Evelyn Roman
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

In this paper we discuss the work being done at Itek combining parallel, symbolic, and neural methodologies at different stages of processing for imagery exploitation. We describe a prototype system we have been implementing combining real-time parallel image processing on an 8-stage parallel image-processing engine (PIPE) computer with expert system software such as our Multi-Sensor Exploitation Assistant system on the Symbolics LISP machine and with neural computations on the PIPE and on its host IBM AT for target recognition and change detection applications. We also provide a summary of basic neural concepts, and show the commonality between neural nets and related mathematics, artificial intelligence, and traditional image processing concepts. This provides us with numerous choices for the implementation of constraint satisfaction, transformational invariance, inference and representational mechanisms, and software lifecycle engineering methodologies in the different computational layers. Our future work may include optical processing as well, for a real-time capability complementing the PIPE's.

Paper Details

Date Published: 1 February 1990
PDF: 12 pages
Proc. SPIE 1156, Airborne Reconnaissance XIII, (1 February 1990); doi: 10.1117/12.962472
Show Author Affiliations
Evelyn Roman, Itek Optical Systems (United States)

Published in SPIE Proceedings Vol. 1156:
Airborne Reconnaissance XIII
Paul A. Henkel; Francis R. LaGesse; Wayne W. Schurter, Editor(s)

© SPIE. Terms of Use
Back to Top