Skip to main content
SLU publication database (SLUpub)

Conference paper2011Peer reviewed

Vegetation classification in the Swedish sub-arctic using a combination of optical satellite images and airborne laser scanner data

Nordkvist, Karin; Nyström, Mattias; Reese, Heather; Holmgren, Johan; Olsson, Håkan

Abstract

The aim of this pilot study was to investigate to which degree the accuracy of automated vegetation classification in the Swedish sub-arctic could be improved by combining optical satellite data with airborne laser scanner (ALS) data, compared to using satellite data only. This information is of interest in an ongoing discussion about the possible inclusion of the mountains in northern Sweden in the national laser scanning that started in 2009. A SPOT 4 scene and ALS data from an Optech ALTM Gemini scanner, both from 2010, were used in maximum likelihood classification. Data for training and validation was obtained from 279 plots with 20 m radius that were visited in field 2010. These plots were located near Abisko in northern Sweden (lat. 68° 23' N, long. 18° 53' E), on the north and south side of Lake Torne Träsk. A classification scheme with 7 classes based on the Swedish mountain vegetation map was used. Classification using only SPOT data gave an over-all accuracy of 75.6%, and the combination of SPOT data and ALS data increased the accuracy to 81.4%

Published in

Title: Silvilaser 2011 - 11th International Conference on LiDAR Applications for Assessing Forest Ecosystems
Publisher: Conference Design Pty Ltd

Conference

Silvilaser 2011: 11th International Conference on LiDAR Applications for Assessing Forest Ecosystems