The goal of the project is to perform feasibility study of land cover classification based on SAR Sentinel-1 images. Currently used algorithms dedicated to land cover classification on SAR images does not give the results on a level of optical images. It tends to take an attempt to resolve this issue. Previous studies indicate that the use of polarimetric technique is the most promising and still not fully recognized. Research work will be focus mainly on identification of built-up areas and other artificial objects as well as on other basic land cover classes: water, forest, bare soil, agriculture, grasslands. If the results allow the number of considered classes will increase. Optical SPOT images will serve as reference data for object identification and evaluation of land cover classification.
The subject of the project is very important for all kinds of research and implied activities, which are related to monitoring of the Earth surface.
Project duration: 01.02.2013 – 31.01.2016