Ghana 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 Ghana. It covers approximately 204,300 kilometers of roads from OpenStreetMap, with deep learning predictions augmenting missing surface data. The dataset distinguishes paved and unpaved roads and is intended for transportation and infrastructure analysis.

Use Cases

• Transportation planning based on paved versus unpaved road classifications.
• Infrastructure analysis using predicted road surface types and lengths.
• Climate emissions assessment leveraging road surface data for vehicle emission models.
• GIS applications integrating road attributes like highway type, surface, and smoothness.

Strengths

• Provides AI-derived surface classifications for approximately 204,300 kilometers of roads.
• Augments OpenStreetMap data with deep learning predictions from Mapillary imagery.
• Includes multiple classification schemes (OSM-priority and DL-priority) for flexibility.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Approximately 144,300 kilometers (70.6%) of road surface information is missing in the original OSM data.

Provenance

Source

OpenStreetMap (OSM), augmented with deep learning predictions from Mapillary imagery and urban layers from GHSU and AFRICAPOLIS.

Collection Method

Surface classification derived using a hybrid deep learning approach, as detailed in the referenced arXiv paper.

Time Range

The mean timestamp of predictions used is provided in the 'DL_mean_timestamp' field.

Freshness

Last updated 2026-03 02 17:20:11.862929; freshness should be verified.

Geography

Ghana