DBPapers
DOI: 10.5593/SGEM2016/B22/S10.097

A MODIFICATION OF KRUSKAL’S ALGORITHM FOR SATELLITE IMAGE SEGMENTATION

M. Pelevin, M. Tsvetkov
Thursday 8 September 2016 by Libadmin2016

References: 16th International Multidisciplinary Scientific GeoConference SGEM 2016, www.sgem.org, SGEM2016 Conference Proceedings, ISBN 978-619-7105-59-9 / ISSN 1314-2704, June 28 - July 6, 2016, Book2 Vol. 2, 759-766 pp

ABSTRACT
Developments in satellite-based scanning systems over the past few decades have led to a remarkable increase in the amount of remote sensing data. With data volumes increasing all the time, it is important to always be mindful of ways to decrease satellite image processing time. One of the most complex problems of image processing that requires an effective solution is the segmentation of original image into homogenous regions.
In this paper we present a modified Kruskal’s algorithm as an effective graph partitioning method for image segmentation. The proposed algorithm creates a tree structure for image description by mapping the tree to the array-based implementation of disjoint sets which are explicitly calculated with Kruskal’s algorithm instead of successive tree forming for different weights of original graph.

Keywords:remote sensing data, image processing, segmentation, Kruskal’s algorithm