Comparison of land cover/usage maps with Sentinel-2 and Landsat-8 satellite images

Land-use data is utilized as the basic information for conducting spatial analysis in the sustainable management of agricultural policies, natural disasters and urban planning. Specifically, in the last decade, remote sensing with quick data access stands out as one of the most effective ways to reach the current information of the land. Good understanding of the earth/land contributes to the production of effective solutions to problems or to the improvement of their current situation. As a result of the literature review in the production of land-use maps; It is seen that the resolution values of the satellite images, the preferred methods during the classification stage and the indices to support the classification are found effective. In this study, the contribution of various plant indices produced by using Sentinel-2 and Landsat-8 satellite images towards image classification accuracy is investigated. The study area was chosen as Bakırçay basin. 7 classes were selected and classified by support vector machines methodology. Classification was carried out with pixel based and controlled classification method. In addition to comparing the general accuracy of satellite images, the accuracy values of 7 different classes discussed in the classification process are examined.