Cuba Road Surface and Passability Data for 12,300 km of Arterial Roads

Mapping 12,300 km of arterial roads in Cuba, this dataset provides AI-derived surface and width attributes generated by HeiGIT from 2020 and 2024 PlanetScope satellite imagery. It fills significant data gaps by providing surface predictions for 12,100 km of road segments that lack tags in OpenStreetMap, focusing on motorway, trunk, primary, and secondary classifications.

Use Cases

• Logistical planning for heavy vehicles using the Road Width Class (1-3) to identify single-lane vs two-lane flow
• Emergency response routing using the Humanitarian Passability Index (HPI) to prioritize accessible routes
• Infrastructure development monitoring by filtering the Road Surface Change column for unpaved-to-paved transitions

Strengths

• 89.2% accuracy for the PlanetScope surface model
• Covers 12,300 km of the national arterial backbone
• Includes temporal change detection between 2020 and 2024
• Standardized 1-6 Humanitarian Passability Index for logistics

Limitations

• Excludes minor, residential, and local road classes
• Relies on AI-derived predictions which may have errors in complex terrain
• Surface tags in OSM were found to be only 64.7% accurate in this region

Provenance

Source

HeiGIT (Heidelberg Institute for Geoinformation Technology)

Collection Method

Deep learning analysis of PlanetScope satellite imagery (3m resolution) integrated with OpenStreetMap road geometries.

Time Range

2020-2024

Freshness

Last updated March 2026; utilizes satellite imagery from 2020 and 2024.

Geography

Cuba