Share Email Print

Proceedings Paper

Accounting For The Atmosphere In Extrapolating System Performance To Different Environments
Author(s): Donald W. Hoock; John C. Giever
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

One goal of system performance modeling is to predict system performance for new, untested environments. When the relationships between environmental variables and system performance variables are clear, the solution reduces to an uncomplicated sensitivity analysis. For complex systems, however, such as a sequence of preprocessing, detection and classification algorithms in an Aided Target Recognition box, the environmental relationships may not be obvious. This paper explores two approaches to including the atmosphere in extrapolating imaging system performance in different environments. One approach is to describe analytically the response of individual image processing steps to atmospheric effects. We consider the sensitivity of an edge detector, target moments, and a basic linear mapping classifier to obscuration, path radiance and turbulence blur. In a second approach one modifies the input images (for example the training images for a classifier) for different weather conditions and then observes the system response to these new inputs. The problems of image modification that remove one set of atmospheric effects while adding new conditions are also discussed.

Paper Details

Date Published: 11 October 1989
PDF: 12 pages
Proc. SPIE 1115, Propagation Engineering, (11 October 1989); doi: 10.1117/12.960883
Show Author Affiliations
Donald W. Hoock, U.S. Army Atmospheric Sciences Laboratory (United States)
John C. Giever, New Mexico State University (United States)

Published in SPIE Proceedings Vol. 1115:
Propagation Engineering
Norman S. Kopeika; Walter B. Miller, Editor(s)

© SPIE. Terms of Use
Back to Top