Global building footprint datasets accuracy evaluated

A new research paper evaluates the accuracy of four prominent global building footprint datasets: Global Human Settlement Layer (GHSL), Microsoft's TEMPO, The Global Building Atlas (GBA), and Overture. Using ORBITaL-Net as ground truth, the study found that GBA and TEMPO generally offered the highest accuracy, though performance varied significantly across different geographic locations, population densities, and income groups. Notably, all assessed products demonstrated reduced accuracy in Africa and Asia, as well as in high-density urban areas. AI

IMPACT This research highlights limitations in AI-driven geospatial data, impacting applications reliant on accurate building footprint information.

RANK_REASON This is a research paper evaluating existing datasets. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.CV →

• Africa
• Global Human Settlement Layer
• Microsoft
• ORBITaL-Net
• Overture Maps Foundation
• Tempo
• The Global Building Atlas

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →