Title :
Separability of Agricultural Cover Types in Spectral Channels and Wavelength Regions
fDate :
7/1/1980 12:00:00 AM
Abstract :
This study was a continuation of a more complete evaluation of the spectral channels as well as wavelength regions-visible, near infrared, middle infrared, and thermal infrared-with respect to their estimated probability of correct classification Pc in discriminating agricultural cover types reported previously by Kumar and Silva [1]. Multispectral scanner data in twelve spectral channels in the wavelength range of 0.4-11.7 ??m acquired in the middle of July for three flightlines were analyzed by applying automatic pattern recognition techniques. The same analysis was performed for the data acquired in the middle of August 1971, over the same three flightlines, to investigate the effect of time on the results. The effect of deletion of each spectral channel as well as each wavelength region on Pc is given. Values of Pc for all possible combinations of wavelength regions in the subsets of one to twelve spectral channels are also given. The overall values of Pc were found to be greater for the data of the middle of August than the data of the middle of July.
Keywords :
Infrared spectra; Pattern analysis; Pattern recognition; Performance analysis; Probability; Space technology; Virtual manufacturing;
Journal_Title :
Geoscience and Remote Sensing, IEEE Transactions on
DOI :
10.1109/TGRS.1980.4307501