LAND COVER FRACTIONS MAPPING USING MODIS IMAGERY OVER NANGGROE ACEH DARUSSALAM (NAD) PROVINCE, INDONESIA

Abstract

Land cover analysis plays an important role in many environmental applications nowadays. The objective of this study is to assess the capability if Moderate Resolution Imaging Spectroradiometer (MODIS) scene for land cover mapping over Nanggroe Aceh Darussalam (NAD) provonce, Indonesia. Standard supervised classification techniques were used in this study namely maximum likelihood, minimum distance-to-mean, parallelepiped and parallelepiped (with Maximum Likelihood Tie Resolution) by using PCI Geomatica 9.1 image processing software. The accuracy of each classification map was assessed using the reference data set consistend of a large number of samples collected per category. The result from this multispectral classification analysis of this study area indicated that urban features could be cleanly identified and classified relative to the surrounding terrain and its associated desert features. In this study, Kappa statistic and overall accuracy were calculated and compared for the three supervised classification technoques. Geometric correction was performed for the digital image using the nearest neighborhood method with second order polynomial. This indicates that land features can be carried out using remote sensing classification method of the MODIS imagery.