1 September 1990 Image metrics approach to understanding effects of terrain and environment on performance of thermal target acquisition systems
Author Affiliations +
Abstract
Understanding the general statistical characteristics and the distribution of target-like features in thermal background imagery is an important part of solving the automatic target recognition (ATR) problem. An alternative approach to test site and scene characterization, based on thermal background image metrics, is described and demonstrated. A database of forward looking infrared (FLIR) imagery, and meteorological and terrain data was systematically obtained from three continental U.S. (CONUS) test sites. Image metrics, relevant to ATR performance, were computed on all imagery. It was deonstrated that temporal variations in these metrics could be predicted (r2 >= 0.79) using current meteorological conditions and a time history of solar loading measurements. Scene-to-scene differences in the texture metrics at a single test site could be predicted (r2 >= 0.78) based on gross scene content attributes. The applications and limitations of this approach and procedure are discussed.
© (1990) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Bruce M. Sabol, Bruce M. Sabol, Kenneth G. Hall, Kenneth G. Hall, } "Image metrics approach to understanding effects of terrain and environment on performance of thermal target acquisition systems", Proc. SPIE 1312, Propagation Engineering: Third in a Series, (1 September 1990); doi: 10.1117/12.21889; https://doi.org/10.1117/12.21889
PROCEEDINGS
20 PAGES


SHARE
Back to Top