North Korea Road Surface and Passability Data for 17,300 km of Arterials

This dataset maps 17,300 km of arterial roads in the Democratic People's Republic of Korea using AI analysis of PlanetScope satellite imagery from 2020 and 2024. Created by the Heidelberg Institute for Geoinformation Technology (HeiGIT), it provides surface types, width classes, and passability scores to supplement OpenStreetMap data. The analysis specifically targets motorway, trunk, primary, and secondary road classifications.

Use Cases

• Monitoring national infrastructure growth by tracking transitions in the 'Road Surface Change' column
• Logistical planning for heavy transport by filtering for 'Road Width Class 3' segments (>5.5 m)
• Assessing regional accessibility for humanitarian aid using 'Humanitarian Passability Index' scores

Strengths

• 89.2% surface prediction accuracy compared to 64.7% in existing OSM tags
• Fills surface data gaps for 15,900 km of roads missing OSM attributes
• Includes temporal change detection between 2020 and 2024 imagery

Limitations

• Geographic restriction to the Democratic People's Republic of Korea
• AI-derived predictions may contain errors inherent to satellite imagery analysis
• Excludes non-arterial road networks which may be critical for last-mile delivery

Provenance

Source

Heidelberg Institute for Geoinformation Technology (HeiGIT)

Collection Method

Deep-learning predictions generated from PlanetScope satellite imagery integrated with OpenStreetMap road geometries.

Time Range

2020-2024

Freshness

Contains imagery-based analysis up to 2024; last metadata update in March 2026.

Geography

Democratic People's Republic of Korea