Python Machine Learning Conference & GeoPython 2020

»Using Google Earth Engine Python API and XGBoost for Built-up classification«
2020-09-21, 11:00–11:30, Room 2

Performing built-up classification using Google Earth Engine python API (for data collection) and XGBoost for classifiers

In this talk, i want to share my experience conducting classification built-up area using Google Earth Engine and XGBoost. The experiment is consist of several steps, as follow: 1. preparing the imagery satelitte data using Google earth engine API (Landsat, VIIRS, and Built-up based imagery) 2. calculating several indices for classification, such as NDVI for vegetation measurement, SAVI for soil observation, MNDWI for watery measurement, NDBI for built measurement, Land Emissivity, and radiance. The first 5 indices are derived from Landsat and the radiance is derived from VIIRS. 3. merge the indices with built-up class from GHSL (a buil-up based imagery as base class) 4. performing the classification in 2015 year (cause the base class is derived in 2015 epoch). the classification is performed using XGBoost 5. using classifiers produced in step 4, perform classification in 2014 and 2019 to get the development of built-up area