Applying Spatial Coherence Method to Landcover Classification with Satellite Imagery

In this study we collected SPOT-5 satellite imagery in northern Taiwan area for classification of land covers using Spatial Coherence Methods (SCM) and validation of the results using Maximum Likelihood Classifier (MLC). One of the advantages using spatial coherence method is that we can quickly filter out most of the surface features of mixture pixels and efficiently obtain the spectral characteristics of the pure features in the study area in order to achieve the purpose of the land covers classification. After completion of the database for the spectral characteristics of land covers, the results would be re-presented by supervised classification in order for combination both the advantages of unsupervised and supervised classification methods.