ANALYSIS OF LAND USE AND LAND COVER CLASSES FOR THE AFFORESTED POST-MINE SITE USING SENTINEL-2 IMAGES.

Autor: Szostak, Marta, Likus-Cieślik, Justyna, Knapik, Kacper, Wężyk, Piotr, Pietrzykowski, Marcin
Předmět:
Zdroj: Proceedings of the International Multidisciplinary Scientific GeoConference SGEM; 2018, Vol. 18, p41-48, 8p
Abstrakt: The study was performed at the former Jeziórko sulphur mine (216.5 ha of afforested area after borehole exploitation). The area was characterized by the terrain structure and vegetation cover resulting from the reclamation process. The applied direction of reclamation for this area was forestry. The study investigated the possibility of applying the latest, freely available Sentinel-2 satellite images for the automation of land use and land cover (LULC) mapping, mainly in terms of monitoring the forest succession process in relation to the mode of reclamation. It was verified whether the processing of Sentinel-2 images allows for reliable LULC classification, mainly in the identification of forest succession areas, in relation to LULC mapping prepared by the photointerpretation and manual vectorization of orthophotomaps. The obtained classification results for Sentinel-2 data were also compared to the results of Landsat 8 image processing. The results of the Sentinel-2 image classification showed a correct graphical representation of the LULC classes in relation to the results from the applied vectorization of aerial orthophotomaps - better than the results of the Landsat 8 image processing. The area of the Forest class, based on the classification of the Sentinel-2 images, compared to the results of the vectorization of the orthophotomaps, showed a 1.8% difference in the percentage of LULC classes in the analysed area (for Landsat 8, there was an 8.8% difference). For the Transitional woodland shrub class, there was a 4.9% difference (Landsat 8: 10.3%). [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index