Thermal data from the Landsat-3 Multispectral Scanner (MSS) are used in conjunction with visible near-infrared data associated with a scene in New York State in order to assess the contribution of thermal data in classification. Despite limitations resulting from poor thermal resolution (about plus/minus 1.5 C), a spatial resolution 1/9 of the other bands, and instrumentation problems, the thermal data are shown by principal component analysis to provide additional information representing a potential discriminating factor for multispectral classification. However, analysis shows that this information is not readily associated with that in the visible near-infrared for identification of surface types and may lead to misinterpretation or spurious classification unless considerable caution is exercised. A physical explanation is suggested involving thermal fluxes at the Earth's surface.
The contribution of thermal data in Landsat multispectral classification
Der Beitrag der thermischen Daten zur multispektralen Klassifikation der Erdoberflaeche mit dem Landsat-3 Satelliten
Photogrammetric Engineering and Remote Sensing ; 47 , 2 ; 229-236
1981
8 Seiten, 7 Bilder, 4 Tabellen, 13 Quellen
Article (Journal)
English
"Landsat Multispectral Data and Introduction to Landsat 6"
British Library Online Contents | 1992
|Multispectral Landsat Image Classification Using Fuzzy Expert Systems
British Library Conference Proceedings | 2004
|Calibration of the thermal band of the Landsat Multispectral Scanner
Tema Archive | 1978
|LANDSAT-3 has multispectral scanner problem
NTRS | 1979
Multispectral linear arrays as an alternative to Landsat-D
Tema Archive | 1979
|