Urban land use classification of São Paulo using Machine Learning and Sentinel 2

Authors

DOI:

https://doi.org/10.11606/rdg.v0ispe.145784

Keywords:

GEE, Machine Learning Classifier, Temporal Index, Texture Index

Abstract

Map Urban and intra-urban is essential for large city planning. Google Earth Engine, machine learning and Sentinel 2 images allow a detailed classification of urban areas that can be improved by the set of bands used, algorithm, and sample balancing. Classifications were produced for municipality of São Paulo in 2017 with the best result produced by Random Forest, with 87.2% global accuracy when using reflectance, spectral, temporal and texture bands. The result demonstrates the great capacity to use free platform and images with machine learning to classify urban and intra-urban areas.

Downloads

Download data is not yet available.

Published

2018-09-09

How to Cite

Rosa, M. R. (2018). Urban land use classification of São Paulo using Machine Learning and Sentinel 2. Revista Do Departamento De Geografia, spe, 15-21. https://doi.org/10.11606/rdg.v0ispe.145784