Image processing to accomplish automatic recognition of military vehicles has promised incrased weapons systems effectiveness and reduced timelines for a number of Department of Defense missions. Testing these ATR subsystems in most cases has been limited to a handful of the scenario conditions of interest, as is represented by imagery collected with the desired imaging sensor. The question naturally arises as to how robust the performance of the ATR is for all scenario conditions of interest, not just for the set of imagery upon which an algorithm was trained. This paper addresses the utility of image characterization measures in terms of estimating ATR detection performance by correlation analyses between nine different image measures and the detection responses of two ATR algorithm. The algorithms analyzed in this study consisted of combined pre-sceener, detector, and statistical classifier components, and the ATR decision categorized the vehicle as either tracked or wheel edvehicle. No model predictions for hypothesized target signatures or trial comparisons between hypothesized and extracted target signatures were employed in these algorithms. The details of the ATRs are not delineated in this paper. The algorithms were trained on infrared imagery containing hot and exercised military vehicles in 'low clutter' backgrounds, and later tested using a set of imagery collected at the same location, times of day, and under the same weather conditions. The results of the tests in terms of detection performance alone (no false alarms) were employed in this analysis.
Image characterization for automatic target recognition algorithm evaluations
Charakterisierung von Bildern für die Bewertung von Algorithmen zur automatischen Zielerkennung
1990
11 Seiten, 6 Bilder, 5 Tabellen, 5 Quellen
Conference paper
English
Automatic Target Recognition Directed Image Compression
AIAA | 1999
|Image Matting for Automatic Target Recognition
IEEE | 2017
|Distributed image processing for automatic target recognition
Tema Archive | 2001
|