Six sequential Landsat MSS scenes for a United Kingdom test site were geometrically corrected. Resampling to 50 m pixel size was carried out with Bilinear interpolation and nearest neighbour. Spectral coincident plots and decision boundaries were used for feature selection, and supervised maximum likelihood classification used for crop classification. Bilinear interpolation gave a mean increase in classification accuracy of 1.95 % over nearest neighbour. Multitemporal data gave better overall classification accuracies than single date images. A spring/early summer combination gave a mean classification purity of 70 %, a 6 % increase over the best single date classification from May, and 46 & better than the worst from February.
The influence of resampling method and multitemporal Landsat imagery on crop classification accuracy in the United Kingdom
Der Einfluss von Resampling Verfahren und multitemporaeren Landsat-Bildern auf die Genauigkeit der Ernteklassifizierung in Grossbritannien
1986
6 Seiten, 1 Bild, 6 Tabellen, 17 Quellen
Conference paper
English
Supervised Crop Classification from Middle-Resolution Multitemporal Images
British Library Conference Proceedings | 2008
|British Library Conference Proceedings | 2006
|Per field filtering and classification of multitemporal ERS-1 SAR imagery
British Library Online Contents | 1993
|Resampling of Scanned Imagery for Rectification and Registration
British Library Conference Proceedings | 1994
|