Cuba Road Surface Classification from OpenStreetMap and AI Predictions

A 2026 dataset from the Heidelberg Institute for Geoinformation Technology (HeiGIT) provides AI-derived road surface classifications for Cuba. It covers approximately 130,200 kilometers of roads from OpenStreetMap, distinguishing paved and unpaved surfaces using a hybrid deep learning method. The data is augmented with predictions from Mapillary imagery and urban classifications from external layers.

Use Cases

• Transportation planning based on paved versus unpaved road classifications.
• Infrastructure analysis leveraging the combined surface classifications from OSM and deep learning.
• Climate emissions assessment using road surface type and urban/rural classification attributes.
• GIS applications integrating road geometry with surface quality and highway type data.

Strengths

• Provides AI-derived estimates for 130,200 kilometers of roads, quantifying paved (15,000 km) and unpaved (35,900 km) segments.
• Includes multiple classification fields (e.g., pred_class, osm_surface_class, combined_surface_DL_priority) for flexible analysis.
• Augments OSM data with deep learning predictions to address missing surface information for 79,400 km of roads.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.
• Approximately 60.96% of road surface information is missing in the original OSM source.

Provenance

Source

Heidelberg Institute for Geoinformation Technology (HeiGIT)

Collection Method

Derived from OpenStreetMap (OSM) and augmented with deep learning predictions based on Mapillary street-level imagery, combined with GHSU and AFRICAPOLIS urban layers.

Time Range

null

Freshness

Last updated 2026-03-02 17:13:02.574119; freshness should be verified.

Geography

Cuba