Incorporation of digital elevation models with Landsat-TM data to improve land cover classification accuracy

Effective management of natural resources and sound decision making require accurate information. Extraction of accurate information from remotely sensed data has been limited by several factors, most importantly the effect of topography, particularly in rugged terrain. Digital elevation models (DEMs) have proved to be an effective aid to improving landcover classification. The principal objective of this study is to evaluate the contribution and quantify the effectiveness of DEMs in improving landcover classification using Landsat-TM data over a rugged area in the Atlas Mountains, Morocco. This study showed that DEM data considerably improved the classification accuracy by reducing the effect of relief on satellite images. The variation coefficient (standard deviation divided by the mean) for homogeneous cover type areas was substantially reduced for all the spectral bands on the corrected image. consequently, the overall accuracy, the Kappa coefficient, and the Tau coefficient were notably improved on the corrected image. The individual accuracies of the different classes also increased by up to 60%. (C) 2000 Elsevier Science B.V. All rights reserved.