Land cover classification using various remote sensing datasets with PlanetScope SuperDove Data in GEE

Abstract

Planet operates the Earth observation satellites PlanetScope (PS) and SkySat (SS). The
PlanetScope satellite constellation consists of multiple launches of groups of satellites.
Consisting of approximately 130 satellites, PlanetScope is capable of examining the entire
Earth's surface every day. The aim of this study is to investigate the effect of Green I (b3 band
513-549nm) and Yellow (b5 band 600-680nm) bands of PlanetScope PSB.SD sensor on
different land cover classes. In addition, the interaction of these bands with various indexbased algorithms calculated from Sentinel-2 satellite images was examined. The results
showed that different datasets had different overall accuracies. This study was carried out
with the Google Earth Engine (GEE), a cloud computing platform with a very comprehensive
database of remote sensor data and satellite images designed for the analysis of geographic
data. By comparing the data sets, the effect of the GreenI and Yellow bands of the Planetscope
Superdove (PSB.SD) sensors in two different study areas, Karataş and Doğubeyazıt districts,
also shows in different classes. As a result, a highly accurate land cover classification was
established in the study areas. The results can be critical in choosing bands and indexes
depending on the study area and class type decided in a land cover classification.