Conventional approaches to inventories from satellites rely heavily on the judgment of image analysts, who 'train' the computer classification program. The fully automated approach to this problem, described in the present paper, eliminates the need for human judgment in training the computer, by using fully automated techniques for selecting spectral characteristics and classification parameters to optimize agreement with ground truth in specified test areas.
Toward fully automated agricultural inventory from satellites
Plenary Meeting ; 1977 ; Tel Aviv, Israel
1977-06-01
Conference paper
No indication
English
Toward a Fully-Automated Digital Holographic Spray Analyzer
British Library Conference Proceedings | 2008
|Fully MMIC Hybrid Equipment for Telecommunication Satellites
British Library Conference Proceedings | 1998
|