DBPapers
DOI: 10.5593/SGEM2014/B23/S10.007

AUTOMATIC CLASSIFICATION OF FOREST STAND BOUNDARIES BASED ON RESULTS OF AERIAL PHOTOGRAMMETRY

J. Solanka, M. Kurcikova, F. Chudy
Wednesday 1 October 2014 by Libadmin2014

References: 14th International Multidisciplinary Scientific GeoConference SGEM 2014, www.sgem.org, SGEM2014 Conference Proceedings, ISBN 978-619-7105-12-4 / ISSN 1314-2704, June 19-25, 2014, Book 2, Vol. 3, 55-62 pp

ABSTRACT
This paper deals with an assessment of accuracy of the automatic object-oriented classification of margins and internal subdivisions of forest stands using different orthophotos. For elaboration, we used outputs of digital photogrammetry - the traditional and true orthophotos made by large format digital camera UltraCam Xp with a final spatial resolution of 10 cm and outputs of small-scale imaging using UAV Gatewing X100 with a spatial resolution of 10 cm. The paper also evaluated the accuracy of aerotriangulation performed during the processing of particular materials, according to the amount of ground control points used. A skilled operator determined the correct boundaries (reference standard) by stereoscopic evaluation, from which a comparative TTA mask was created. The classification accuracy was evaluated with the KHAT index calculated on the basis of the created TTA mask. Our results have shown that in case of using the object-oriented classification for identifying the boundaries of forest stands from orthophotos we have achieved satisfactory results not only in terms of accuracy, but also in terms of efficiency and time demands. The best classification accuracy can be achieved if true orthophotos are used. The research site was located in the territory of the University Forest Enterprise at the Technical University in Zvolen, in the central part of Slovakia. The selected sample area (possibility of using UAVs was taken into account) is situated in a slightly hilly terrain with forests and herbal vegetation cover. The area consists of compact forest stands, forests vertically subdivided by anthropogenic interactions according to forest management rules of silviculture, as well as a significant boundary between forest and the agricultural land.

Keywords: object oriented classification, aerial digital photogrammetry, aerotriangulation, identification of forest stand