DETECTION OF LAND COVER CHANGES USING MULTI-TEMPORAL SATELLITE IMAGERY

No Thumbnail Available
Date
2015
Authors
Elfaki, Elnour
Journal Title
Journal ISSN
Volume Title
Publisher
University of Khartoum
Abstract
In this study, two different change detection techniques were applied in order to assess land cover changes in El Rawashda forest, Sudan: comparison of classification and multivariate alteration detection. Firstly, two satellite imagery, acquired in 2003 by Landsat ETM+ and by ASTER in 2006, were classified into four main land cover classes namely grass land, close forest, open forest and bare land. A change matrix was created in order to map the land cover changes from 2003 to 2006. Generally, the results show a noticeable increase in area on both close forest and open forest areas with decrease in grass lands within the period 2003-2006. More than one third of grassland (36%) was converted to close forest, one fourth (24%) to open forest areas. In the three-year period, 9079 hectares of open forest, (8% of the investigation area), were transformed to close forest. Secondly, we perform linear transformation by applying the multivariate alteration detection (MAD), the MAD components were then examined to identify the quality of changes.
Description
Keywords
Multi-temporal Data, Change Detection, Pixel-based Techniques, Land Cover Changes.
Citation