Author : Willem Anthon Siahaya*
The objective of this research are (a) determine land cover type based on NDVI and (b) RVI transformation model on landsat image, (c) land cover type determination using the threshold model approach. The research method used is Landsat TM image interpretation digitally by using spectral value transformation (Normalized Differenced Vegetation Index/NDVI) and determination of land cover type using threshold model approach and field check. The results indicated there are five types of land cover, ie (i) those are (i) bare land and waterproof surface, (ii) herbaceous and shrubs, (iii) low density vegetation, (iv) Medium density vegetation and (v) High density vegetation. There is a significant correlation between RVI and NDVI, with vegetation density, this is indicated by r2 values of 0.809 and 0.948 respectively. Transformation of RVI and NDVI on Landsat TM5 with threshold model approach can be used to obtain land cover information with relatively high accuracy. The accuracy test showed overall accuracy of NDVI was 93.0%, while the overall accuracy of RVI was 88.9%. Key words: Land Cover, Vegetation Density, Vegetation Indices, Landsat Imagery.
Dr. Arend L Mapanawang, Sp.PD, FINASIM, PhD