Determination of CORINE Land Cover/Land Use Classes Using Satellite Images with Different Spatial Resolution

Abstract

Determining the current spatial distribution of land cover/use classes and examining the changes occurring in the process constitutes an important basis for studies conducted in many economic and socio-cultural areas. Therefore, for the information on land cover/land use to be consistent among themselves, it must be systematically classified and produced to certain standards. The aim of this study was aimed to produce distribution maps of the land use/land cover classes from Landsat 8, Sentinel 2, and Triplesat satellite images of an area of 11251 hectare in Samsun, Vezirköprü (Turkey) district and to compare them with the ground reality. According to the results obtained, it was determined that the most widely distributed class in all satellite images of the study area was the agricultural areas. In terms of classification accuracy, Kappa values (92.95% and 93.11%) of Sentinel and Triplesat satellite images were proportionally similar, while this value decreased to 83% in the Landsat satellite image. In addition, it has been determined that some of the agricultural and forest areas have shifted to artificial areas in a 34-year period. As a result of the study, it has been suggested that reliable results obtained in the monitoring of land cover/land use can benefit from the ability to observe large areas of satellite images in short periods and high resolution.

Keywords

• corine
• satellite imagery
• samsun
• spatial distribution
• image processing