Mapping Urban Land Cover by Using Landsat, Sentinel Images and Open Social Data

摘要

Mapping urban land cover is one of the important issues in urban planning of the developing country. This paper investigates a methodology to classify land cover of Yangon by using the web-based remote sensing platform, Google Earth Engine (GEE). Features are extracted from 30m Landsat-8 images by using pixel-based approach and 10m Sentinel-2 images by using object-based approach for land cover classification. In addition, open social data extracted from Open Street Map (OSM) and ground truth data are applied as additional features to enhance the accuracy of the classifier.