Separability of Mangrove Landcover by Using RapidEye Satellite Imagery

Abstract

RapidEye satellite imagery was used to classify two schemes of coastal landcover in Liong River, Bengkalis Island. Transformed divergence were used due to separability analysis and accuracy of convergence matrices was used kappa’s test. Two landcover classification schemes were used. The first scheme consist of 8 classes and 26 band combination of RapidEye classification result were not promised, especially when mangrove class paired with the forest and rubber plantation classes. The other scheme shownthe better solution since it has three classes (mangrove, waters body and others).