The methodology proposed for the development of the CopernicusLAC Chile Urban Atlas is based on the identification of five categories of land uses present in the capitals of the member countries. The first version of the Atlas (at a resolution of 10 metres) aims to represent five urban land uses in a level 2 classification based on spectral characteristics. The proposed classes to be detected are Buildings, Transport Units, Urban Vegetation, Bare Soil and Water Bodies.

The present classification is based on machine learning models using Sentinel-2 satellite products with cloud-free temporal composition using their spectral bands at 10m resolution. Moreover, the in-situ data were complemented with auxiliary open access information on urban infrastructure and equipment.

This comprehensive, data-driven approach will enable the generation of a robust and useful Urban Atlas for local authorities, urban planners and other stakeholders interested in the sustainable development of cities in Latin America and the Caribbean.