NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Back to Results
Landsat Thematic Mapper digital information content for agricultural environmentsLandsat Thematic Mapper (TM) data collected for Imperial Valley, California in December, 1982 were digitally examined to assess their utility to distinguish among agricultural and other land-covers. Statistics for thirty-seven training sites representing a variety of crops plus urban, water and desert land-covers were obtained and analyzed using transformed divergence (TD) calculations. TD values were employed to assess intraclass variability and the best bands for classification. Four subscenes were selected for clustering or unsupervised signature extraction. These areas were agriculture, urban, desert and water land-covers. The number of clusters for these subscenes were examined and the best TM bands for interclass separability were identified. The results of the clustering and training site analyses for interclass separability were compared. The TM data were useful for the digital delimitation of most crops and other cover types in this analysis. Four bands of data are adequate for classification with the best results obtained by the selection of one band from each of the available portions of the electromagnetic spectrum. Different band combinations are best for various land-cover intraclass separability.
Document ID
19870036556
Document Type
Conference Paper
Authors
Haack, Barry (George Mason University Fairfax, VA, United States)
Bryant, Nevin (George Mason Univ. Fairfax, VA, United States)
Adams, Steven (California Institute of Technology Jet Propulsion Laboratory, Pasadena, United States)
Date Acquired
August 13, 2013
Publication Date
January 1, 1986
Subject Category
EARTH RESOURCES AND REMOTE SENSING
Funding Number(s)
CONTRACT_GRANT: NAS7-918
Distribution Limits
Public
Copyright
Other