Abstract: The aim of this study is to conduct a forest area classification using optical and
synthetic aperture radar (SAR) satellite data sets. For this purpose, a forested site around
the Lake Khuvsgul situated in northern Mongolia is selected. As remote sensing (RS) data
sources, multispectral Landsat 8 images and ALOS-2 PALSAR L-band HH polarization
data are used. To produce a reliable land cover map from the combined multichannel
and SAR images, a support vector machine (SVM) and maximum likelihood classification
(MLC) techniques are compared. For the accuracy assessment, an overall accuracy that
applies a confusion matrix is used. Overall, the research demonstrates that the integrated
optical and microwave RS can be efficiently used for the forest area discrimination.
Keywords: Optical, Microwave, Forest area, Classification
