Extraction and Validation of Database of Urban and Non-Urban Points from Remote Sensing Data

Recent advancements in global settlement mapping have ushered in a new era of spatial analysis.The introduction of the Global Urban Footprint (GUF)-derived from radar-based TanDEM-X data-and the Global Human Settlement Layer (GHSL)based on optical satellite data-offers unprecedented spatial resolution for mapping complex human settlement patterns.Despite these innovations, comparative assessments evaluating their performance against existing lower-resolution datasets remain limited.To address this, we present a robust cross-comparison framework that examines inter-map agreement and classification accuracy across multiple African landscapes with varying settlement typologies.Using a curated set of reference data points from thirteen African countries, the study assesses the concordance between ground observations and GUF outputs.Key indicators of accuracy, including the Kappa coefficient, overall agreement rates, and pattern-based metrics, are used to evaluate the performance of these new layers.Findings reveal that the high-resolution GUF and GHSL products offer significant improvements in delineating low-density and periurban settlements-areas previously underrepresented by coarser-resolution maps.Pattern-based analysis shows that accuracy is positively correlated with settlement structure, particularly in zones characterized by medium and small patch sizes.The results indicate a marked shift in the spatial fidelity of global settlement representations, particularly beyond urban core areas.This comprehensive validation supports the integration of GUF and GHSL data into urban planning and policy frameworks, enhancing the capacity to monitor and manage urbanization at multiple scales.