Guinea-Bissau Road Surface, Width, and Passability Data (2020-2024)

HeiGIT generated this dataset of AI-derived road attributes for approximately 900 km of arterial roads in Guinea-Bissau using PlanetScope satellite imagery from 2020 and 2024. It provides surface type, width classifications, and humanitarian passability scores for segments classified as motorway, trunk, primary, and secondary in OpenStreetMap.

Use Cases

• Evaluating logistical accessibility for humanitarian aid using the Humanitarian Passability Index (HPI) scores
• Monitoring infrastructure development by tracking transitions in the Road Surface Change column
• Determining heavy vehicle route feasibility based on the Road Width Class (1-3) categories

Strengths

• 89.2% accuracy for surface predictions, exceeding the 64.7% accuracy of OSM surface tags
• Multi-temporal coverage (2020 and 2024) enables longitudinal analysis of road improvements
• Fills surface tag gaps for 100% of arterial roads in the study area

Limitations

• Excludes tertiary and local road networks
• Small geographic scale with approximately 900 km of total road length
• Potential AI classification noise from dense canopy cover or cloud interference

Provenance

Source

Heidelberg Institute for Geoinformation Technology (HeiGIT)

Collection Method

Deep learning models applied to PlanetScope satellite imagery and integrated with OpenStreetMap road network data.

Time Range

2020-2024

Freshness

Last updated March 2026; imagery analysis covers up to 2024.

Geography

Guinea-Bissau