Inter-sensor comparison of built-up derived from Landsat, Sentinel-1, Sentinel-2 and SPOT5/SPOT6 over selected cities

In the last 5 years, several information layers describing human settlements were developed within the Global Human Settlement infrastructure of the Joint Research Centre using Earth Observation data. Each layer was derived from a different satellite (with different various spatial resolutions and radiometric properties) and from images acquired at different time stamps. The next step is to exploit the synergies between the different sensors and possibly integrate the information layers within a single product. To enable those future developments, it is essential to understand the potentials and limitations of each of the layers and identify complementarities. In these regards, the validation of built-up derived from different sensors is crucial for gaining a deeper understanding of the consistency and interoperability between them. This report, presents the methodology and the results of the inter-sensor comparison of built-up derived from Landsat, Sentinel-1, Sentinel-2 and SPOT5/SPOT6. The assessment was performed for 13 cities across the world for which fine scale reference building footprints were available. Several validation approaches were used: cumulative built-up curve analysis, pixel by pixel performance metrics and regression analysis. The results indicate that the Sentinel-1 and Sentinel-2 highly contribute to the improved built-up detection compared to Landsat. However, Sentinel-2 tends to show high omission errors while Landsat tends to have the lowest omission error. The built-up obtained from SPOT5/SPOT6 show high consistency with the reference data for all European cities and hence can be potentially considered as a reference dataset for wall-to-wall validation at the European level. It was noted that the validation results can highly vary across all the study sites because of the different landscapes, settlement structures and densities.