%0 Journal Article
%T Generalization of heterogeneous alpine vegetation in air photo-based image classification, Latnjajaure catchment, northern Sweden
%A Lindblad
%A K. E. M.
%A Nyberg
%A R.
%A Molau
%A U.
%J Pirineos : Revista de Ecolog赤a de Montaˋa
%D 2006
%I Consejo Superior de Investigaciones Cient赤ficas
%X Mapping alpine vegetation at a meso-scale (catchment level) using remote sensing presents difficulties due to a patchy distribution and heterogeneous spectral appearance of the plant cover. We discuss issues of generalization and accuracy assessment in this case study when using a digital CIR air photo for an automatic classification of the dominant plant communities. Spectral information from an aerial photograph was supplemented by classified plant communities in field and by topographical information derived from a DEM. 150 control points were tracked in the field using a GPS. The outcome from three alternative classifications was analysed by Kappa statistics, user＊s and producer＊s accuracy. Overall accuracy did not differ between the classifications although producer＊s and user＊s accuracy for separate classes differed together with total surface (ha) and distribution. Manual accuracy assessment when recording the occurrence of the correct class within a radius of 5 meters from the control points generated an improvement of 16 % of the total accuracy. About 10 plant communities could be classified with acceptable accuracy where the chosen classification scheme determined the final outcome. If a high resolution pixel mosaic is generalized to units that match the positional accuracy of simple GPS this generalization may also influence the information content of the image. Hemos llevado a cabo la cartograf赤a de la vegetaci車n alpina a escala media (nivel de cuenca experimental) mediante interpretaci車n remota. Esta metodolog赤a plantea dificultades debido a la distribuci車n en mosaico de la vegetaci車n y a la heterogeneidad del espetro obtenido. Se discuten las posibilidades de generalizaci車n de los resultados y el grado de precisi車n alcanzado en este caso experimental mediante fotograf赤a a谷rea digital CIR aplicada a una clasificaci車n autom芍tica de las comunidades vegetales dominantes. La informaci車n espectral obtenida por foto 芍erea se complement車 con la clasificaci車n de las comunidades vegetales in situ y la informaci車n topogr芍fica derivada de un Modelo Digital de Terreno. Adem芍s se marcaron 150 puntos de control en el campo por medio de GPS. Los resultados de tres clasificaciones alternativas se analizaron mediante el estad赤stico Kappa y la precisi車n del usuario y del productor. El grado de precisi車n obtenido apenas difiri車 entre clasificaciones, a pesar de que s赤 hab赤a diferencias significativas entre la precisi車n del usuario o del productor para las diferentes clases, as赤 como para la superficie total y la distribuci車n. La presencia sobre el terreno de
%K Tundra
%K plant community
%K GIS
%K CIR-air photo
%K digital classification
%K Tundra
%K comunidades vegetales
%K GIS
%K fotos a谷reas
%K CIR
%K clasificaci車n
%U http://pirineos.revistas.csic.es/index.php/pirineos/article/view/1/1