Share Email Print

Proceedings Paper

ROCSAT-2 spectral reflectance experiments using local three-end-member spectral mixed model in Taiwan
Author(s): Chih-Li Chang; Chi-Nan Wu; Tzu-Yi Liao; Chiuder Hsiao
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Classification of land cover and land use is one of ROCSAT-2 primary applications. Spectral reflectance experiments were conducted for classification of the main land use and land cover in Taiwan. The four Level I categories of USGS classification system were picked out for classification of the major territory in Taiwan. The fourteen subcategories with fourteen different crops were further classified for the agricultural land use. The barren land is classified into five subcategories. The spectral reflectance of each category or subcategory was recorded with a spectro-radiometer, GER3700, with 702 channels. Since the spectral range of GER3700 covers the four color bands, blue, green, red and infrared of the ROCSAT-2 Remote Sensing Imager, the recorded data were derived into simulated reflectance in the color bands of the ROCSAT-2 Remote Sensing Imager. To use a multiple end-member spectral mixed model, the correlation among the category and/or subcategory was calculated. Yam, dry sand and water were selected as the local three end-members for their low correlation of spectral reflectance. The spectral reflectance of other categories is coded as the components of the local end-members.

Paper Details

Date Published: 17 March 2003
PDF: 8 pages
Proc. SPIE 4879, Remote Sensing for Agriculture, Ecosystems, and Hydrology IV, (17 March 2003); doi: 10.1117/12.462384
Show Author Affiliations
Chih-Li Chang, National Space Program Office (United States)
Chi-Nan Wu, Industrial Technology Research Institute (United States)
Tzu-Yi Liao, Industrial Technology Research Institute (United States)
Chiuder Hsiao, National Space Program Office (Taiwan)

Published in SPIE Proceedings Vol. 4879:
Remote Sensing for Agriculture, Ecosystems, and Hydrology IV
Manfred Owe; Guido D'Urso; Leonidas Toulios, Editor(s)

© SPIE. Terms of Use
Back to Top