EARSeL eProceedings Vol. 3, No. 2, 191-207, 2004

Abstract
Remote sensing can give information about the configuration and composition of coral reefs, about the biophysical parameters of the seas and oceans in which they occur and about the changes over time of these elements. This paper deals with the classification of a Landsat7 ETM+ data set in order to identify the different bottom types (macro-algae, coral, sea grass and sand) occurring on the reefs offshore Hurghada, Egypt. Before classification, the radiance values received at sensor are corrected for atmospheric and water column effects. 'Depth-invariant bottom indices' are calculated and form the basis for classification. Besides the bottom type as an ecological classification, also a geomorphological classification is made. After contextual editing of the ecological classification, both results are combined into an open-ended hierarchical classification scheme. An in-depth accuracy assessment still needs to be undertaken but a mean accuracy between 47% and 83% is to be expected.

ISSN 1729-3782