Klasifikasi Wilayah Daratan, Vegetasi dan Perairan, Menggunakan SVM Serta Penghitungan Luas Wilayah dengan Pixel Klasifikasi Melalui Citra Satelit Landsat-8

Abstract

Area measurement of Land territory with manual measurement by visiting the area to be measured quite tiring and waste energy. Measurement through Google Earth are more effiecient as it doesn’t consume any stamina, however measurement doesn’t classify about the area that are measured. Measurement via Qgis and Arcgis can be done, however it takes a long time to finish it because the data need to be classified earlier on the area to be measured and can only be done on one time only. This research. Uses SVM method which are used on model as much as 415.930 row of data which are trained to be able to predict the area size according to the data which are mentioned before. The research using SVM in predicting area size with a few category which are City, Vegetation, Standing Vegetation, Waters, inland water measured by utilizing the use of Semantic Segmentation from which uses the pixel provided by the imagery of measurement with the purpose of displaying perfect imagery of the measurement which have a quite high accuracy around 95%.