Geospatial Machine Learning in Urban Environments: Challenges and Prospects

Abstract

Accurate and current land cover information is required to develop strategies for sustainable development and to improve quality of life in urban areas. The past decades has seen an increased availability of earth observation satellite (EOS) sensors (e.g., Sentinel-1 and Sentinel-2) as well as machine learning (ML) techniques (support vector machines, random forests) for land cover mapping. While significant progress has made to improve land cover mapping in urban areas, challenges still remain. The purpose of this chapter is to discuss briefly about geospatial machine learning in urban environments as well as some of its major challenges and prospects. The chapter will cover an introduction to geospatial ML (remote sensing image pre-processing and ML techniques), study area and data sets, hands-on exercises, summary, and additional exercises.