Latvia: Road Surface, Width, and Passability for 20,600 km of Arterial Roads

HeiGIT produced this dataset covering approximately 20,600 km of arterial roads in Latvia using PlanetScope satellite imagery from 2020 and 2024. It provides AI-derived classifications for road surface, width, and passability for segments classified in OpenStreetMap as motorway, trunk, primary, and secondary roads.

Use Cases

• Monitoring infrastructure development by identifying transitions from unpaved to paved in the 'Road Surface Change' column
• Logistics and heavy vehicle routing using 'Road Width Class' to identify segments wider than 5.5 meters
• Humanitarian accessibility mapping using the 'Humanitarian Passability Index (HPI)' to assess transport risk

Strengths

• Surface model achieves 89.2% accuracy compared to 64.7% for standard OSM surface tags
• Provides temporal change detection for road paving between 2020 and 2024
• Includes a standardized Humanitarian Passability Index (HPI) for logistical planning

Limitations

• Analysis is restricted to arterial roads and excludes the local/residential road network
• Road width is categorized into three broad classes rather than precise meter measurements
• AI predictions may be affected by satellite imagery artifacts or environmental obstructions

Provenance

Source

Heidelberg Institute for Geoinformation Technology (HeiGIT)

Collection Method

Deep learning analysis of PlanetScope satellite imagery integrated with OpenStreetMap road geometries

Time Range

2020-2024

Freshness

Last updated March 2026; contains imagery-derived data from 2020 and 2024.

Geography

Latvia