inscription
PORTAIL D'INFORMATION GÉOGRAPHIQUE

A geostatistically weighted k-NN classifier for remotely sensed imagery

Geostatistical methods in geography part II : applications in physical geography. Special issue

Auteur(s) et Affiliation(s)

ATKINSON, P.M.
School of Geography, Univ., Southampton, Royaume-Uni
NASER, D.K.
School of Geography, Univ., Southampton, Royaume-Uni


Description :
This study aims to increase the accuracy with which remotely sensed data can be used to generate thematic maps of land cover classes. It explores the use of geostatisical models to characterize the inherent spatial variation between different land covers (woodland, rough grassland, managed grassland, and built land) and integrates these into a supervised, nonparametric, k-nearest neighbor (k-NN) per-pixel classifier. The increase in accuracy obtained by incorporating the geographical weighting is assessed empirically using a spatially and spectrally variable IKONOS subscene.


Type de document :
Article de monographie

Source :
Geographical analysis, issn : 0016-7363, 2010, vol. 42, n°. 2, p. 204-225, nombre de pages : 22, Références bibliographiques : 3 p.

Date :
2010

Editeur :
Pays édition : Etats-Unis, Columbus, OH, Ohio State University Press

Langue :
Anglais
Droits :
Tous droits réservés © Prodig - Bibliographie Géographique Internationale (BGI)