(Article of periodic en Anglais - 2010)

Document title

A geostatistically weighted k-NN classifier for remotely sensed imagery

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

Authors(s) and Affiliation(s)

ATKINSON P.M. (1) ; NASER D.K. (1) ;
(1) School of Geography, Univ., Southampton, ROYAUME-UNI


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


Thematical fascicle

published at : Geographical analysis / ISSN 0016-7363

Editor : Ohio State University Press, Columbus, OH - ETATS-UNIS (1969)

Millesime : 2010, vol. 42, no 2 [pp. 204-225]

Bibliographic references : 3 p.

Collation : fig., tabl.



INIST-CNRS, Cote INIST : 17078

Tous droits réservés © Prodig - Bibliographie Géographique Internationale (BGI), 2010
Refdoc record number (ud4) : 23380043 : Permanent link - XML version
Powered by Pxxo