Akademik Veri Yönetim Sistemi
9th International Conference on Recent Advances in Space Technologies (RAST), İstanbul, Türkiye, 11 - 14 Haziran 2019, ss.833-838
The present study aims to evaluate the land cover/use resource for high-resolution map production in a complex landscape mosaic in the central Italy. The classification methodologies are based on manual and supervised and rule-based classification of Gokturk-1 very high resolution imagery. For the different classifications and for results validation the Italian land use inventory (IUTI) database were used. General spatial distributions of selected classes are successfully mapped with both classification approaches. Pixel-based classification results include some salt and pepper effects and spectral mixtures but still provides good delineation of the selected classes. Incorporating image indices, topological and textural functions into object-oriented classification improved the classification accuracy. However, the scale parameter is important to capture the detail levels obtained with Gokturk-1 images. Moreover, considering the high spatial resolution of Gokturk-1 images, photointerpretation approach is also very suitable to identify objects specifically if more detailed classes will be considered. High quality spatial, spectral and radiometric characteristics of the Gokturk-1 satellite makes it possible to create thematically and geometrically detailed geo-spatial information for every local public administration or private sector without spatio-temporal limitations.