Road Surface Classification for Moldova from OpenStreetMap and AI

A 2026 dataset from the Heidelberg Institute for Geoinformation Technology (HeiGIT) provides AI-derived road surface classifications for the Republic of Moldova. It covers approximately 95,500 kilometers of roads from OpenStreetMap, distinguishing paved and unpaved surfaces using a hybrid deep learning model. The data is augmented with predictions from Mapillary imagery and urban classifications from GHSU and AFRICAPOLIS layers.

Use Cases

• Transportation planning based on paved versus unpaved road surface classifications.
• Infrastructure analysis using attributes like highway type, width, lanes, and bridge status.
• Climate emissions assessment based on road surface type and urban/rural classification.
• GIS applications for routing and accessibility studies using the geospatial road network.

Strengths

• Includes AI-derived surface predictions for approximately 95,500 km of roads.
• Combines multiple data sources: OpenStreetMap, Mapillary imagery, and urban settlement layers.
• Provides detailed OSM attributes like highway, surface, smoothness, width, and lanes.
• Offers multiple classification fields (OSM, DL-prioritized, combined) for comparison.

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 79.2% of road surface information is reported as missing in the original OSM data.

Provenance

Source

Heidelberg Institute for Geoinformation Technology (HeiGIT).

Collection Method

Derived from OpenStreetMap and augmented with deep learning predictions based on Mapillary street-level imagery.

Time Range

Mean timestamp of predictions (DL_mean_timestamp) is provided per feature, but overall temporal coverage is not specified.

Freshness

Last updated 2026-03-02 17:21:08.145717; freshness should be verified.

Geography

Republic of Moldova.