Vol. 4, No. 1, 70-78, 2005

Mapping high mountain vegetation using hyperspectral data
Bogdan Zagajewski, Anna Kozlowska, Malgorzata Krowczynska, Marcin Sobczak and Magdalena Wrzesien

The paper presents the methods and first results of vegetation mapping, using field and hyperspectral airborne data in high mountain ecosystems. The research also aims at a comparison of different remote sensing methods of vegetation classification and at creating a map of actual vegetation.

The study was carried on in the Tatra National Park - encompassing subalpine, alpine and subnival belts of the Tatra Mountains. The results of the ground mapping and different image classification approaches were compared.

Maximum likelihood classification method is widely used in many remote sensing applications and can be considered one of the most popular and reliable techniques. Neural network classification is based on training during a training phase, and the proper classification. The training process is based on determining the neural connection weights to make the output signal from the network as close as possible to the expected result. One of the goals was to verify the usefulness of neural networks for classification and to obtain the best results in vegetation recognising using airborne hyperspectral imagery.

For validation of the DAIS and ROSIS image classifications, a detailed large-scale vegetation map was prepared, using traditional field mapping.

View Full Text (pdf file, 700 kB) previous page
Submitted: 22 July 2004
Revised: 13 December 2004
Accepted: 18 Febuary 2005

Zagajewski B, A Kozlowska, M Krowczynska, M Sobczak & M Wrzesien, 2005. Mapping high mountain vegetation using hyperspectral data. EARSeL eProceedings, 4(1), 70-78


EARSeL European Association of Remote Sensing Laboratories, Strasbourg, France


BIS Library and Information System, Carl von Ossietzky University of Oldenburg


ISSN 1729-3782