PlanetScope contributions compared to Sentinel-2, and Landsat-8 for LULC mapping

Up-to-date and accurate land use and land cover (LULC) maps are vital for monitoring various environmental and natural resources management. In this study, we evaluate classification performance of three optical satellite data in an area located in northwestern Morocco. Images from Landsat-8, Sentinel-2, and PlanetScope were exploited between 2020 and 2021. Then, twelve different multitemporal/multisensor combinations were evaluated using overall accuracy, Cohen's kappa, and F-score. A supervised classification was carried out using random forest algorithm. The study area was classified into two levels, including twelve and sixteen land use/ cover classes. These results suggest that when spatial and spectral resolution increase, classification accuracy improves. PlanetScope's LULC classification (overall accuracy > 97%) performs better than Landsat-8 and Sentinel-2 data. In addition, combining three sensors did not significantly improve the overall classification with regards to PlanetScope only. The high-resolution obtained map can be exploited as input for environmental modeling and could help decision makers for sustainable land and ecosystem management.