dc.description.abstract | Landsat 8 satellite was developed by National Aeronautics and Space Administration (NASA) for natural resource monitoring activities. Landsat 8 is a continuation of the previous Landsat missions (Landsat 7) which has damaged since 2003 with the addition in number of channels and wavelength shift for each channel. The aims of this research are to determine the amount and type of land cover classes that can be identified using Landsat 8 imagery based on separability analysis and also to determine the best classification method in land cover mapping classes in Ciamis district. The using method in this research comprised of identification of land cover object that found in the field, training area, separability analysis to obtain the number of class that can be identified based on the digital number, supervised classification to classify land cover separability analysis result then accuracy test and compare its result with the visual image interpretation. Based on separability analysis, the number of classes that can be identified in Ciamis district consist 12 classes which are cloud, shadows of clouds, mangrove forest, rubber forest, coconut grove, rice field, shrub, open land, water bodie, dry land farming, residential and forest/ mixed farm. The best method to classify land cover in Ciamis district is Support Vector Machine with linear function because it has the highest value of kappa accuracy of 76,56% and has a high conformity by the visual image interpretation. | en |