University of Twente Student Theses


Validating Landscape Heterogeneity Mapped by Hyper –Temporal NDVI Images through Line – Transect Data

Berhe, Rishan Araya (2010) Validating Landscape Heterogeneity Mapped by Hyper –Temporal NDVI Images through Line – Transect Data.

[img] PDF
Abstract:Landscape heterogeneity mapped using hyper-temporal NDVI images can give efficient, reliable and effective information at national and regional level for sustainable development plans. The main objective of this study was to validate landscape heterogeneity mapped using hyper temporal NDVI images though transect sampling method. SPOT-vegetation NDVI image data layers for ten years (April 1998 – March 2007) were stacked and classified using ISODATA; of which optimum number of classes was identified. All the classified images up to the best classified classes were superimposed on top of each other to produce the landscape heterogeneity map. Data was collected from transects that were placed randomly in each survey site. High resolution Google images were digitized and described by using lookup table created using field data. Cover percentages of each pixel was estimated by crossing digitized map with NDVI classes. Finally, estimated cover percents of each cover type, NDVI value extracted from stacked image and boundary heterogeneity value extracted from landscape heterogeneity map were plotted for visual analysis. Using stepwise least squares dummy variable regression method NDVI classes were tested for their significant differences in each survey site. This research found that NDVI classes in the heterogeneity map have significant differences using cover percentages of transects. Strong boundaries were also validated using linear regression analysis between the heterogeneity value of pixels and tree cover percentages in transects. Tree cover percent was the main indicative parameter for strong boundaries between map units. Mainly NDVI classes having strong boundaries were found significantly different using tree cover percentages. Landscape heterogeneity map is found powerful to explain the landscape heterogeneity.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page