EARSeL eProceedings Vol. 3, No. 2, 143-153, 2004
Remko de Lange, Mark van Til and Stephen Dury
Abstract
Citation EARSeL European Association of Remote Sensing Laboratories, Paris, France BIS-Verlag
Vegetation
monitoring is an important tool in the evaluation of nature management in the
coastal zone of The Netherlands. Remote sensing images are valuable in order to
investigate spatio-temporal changes in the vegetation. A classification method
has been developed based on airborne hyperspectral data, which were acquired
using the GER EPS-A scanner. A supervised classification method has been used
applying the Spectral Angle Mapper, in combination with an expert system. The
SAM algorithm determines the similarity between spectra of different vegetation
types. The expert system adds extra information about environmental conditions
to the classification in order to improve the discrimination of vegetation
types, which are otherwise spectrally difficult to identify. In our case, this
method gives the opportunity for rapid classification with an overall accuracy
of 60-70%.
View Full Text (pdf file, 500 kB)
History
Submitted: 07 June 2003
Revised: 31 January 2004
Accepted: 23 February 2004
de Lange R, M van Til & S Dury, 2004. The use of hyperspectral data in coastal zone vegetation monitoring. EARSeL eProceedings 3(2), 143-153
BIS Library and Information System, Carl von Ossietzky University of Oldenburg
ISSN 1729-3782