Ar4-50 Model, the Extractor of Spectral Values Into Remote Sensing Image Data-based Land Use Class

Akhbar Akhbar • Muhammad Basir • Bunga Elim Somba • Golar Golar
Journal article Agrivita Journal of Agricultural Science • 2013


This study attempted to develop an extraction model of spectral values ​​of land objects into land use/land cover classes on remote sensing image in the provision of land database for planning, evaluation, and monitoring in agriculture and forestry. This study employed an Isodata method and Knowledge-Based Systems (KBS) using the Landsat 7 ETM+ image in the coverage area of ​​117,799.06  ha, and the SPOT 5 XS image in the coverage area of ​​113,241.37 ha in Palu, Sigi and Donggala. The study found two image models labelled as AR4-50 and SBP-AR4-50. The separability image AR4-50 model has an average capability for separating land object pixels which are statistically 1811.98 to 1972.08 (moderate-good), with the class accuracy of land use/land cover using the image homogeneity model of SBP-AR4-50, which is totally (confusion matrix) 72.15% -87.17%, the accuracy level of land map generator for agricultural land/forestry is in good-excellent category on the Landsat 7 ETM+ and SPOT 5 XS images.




Agrivita Journal of Agricultural Science

Agrivita Journal of Agricultural Science is a peer-reviewed quarterly journal that disseminates o... see more