Mongolia Road Surface Classification from OpenStreetMap and AI Predictions

Mongolia's road network, covering approximately 0.247 million kilometers mapped in OpenStreetMap, is classified by surface type. The dataset, created by HeiGIT and last updated in March 2026, uses a hybrid deep learning approach to distinguish paved from unpaved roads, augmenting OSM data with predictions from Mapillary imagery.

Use Cases

• Transportation planning based on road surface type and length statistics.
• Infrastructure analysis using attributes like highway type, surface, and smoothness.
• Climate emissions assessment leveraging road surface classification for vehicle emission modeling.
• GIS applications utilizing the geospatial road network data with urban/rural classification.

Strengths

• Includes AI-derived surface classification for approximately 0.247 million km of roads.
• Provides specific statistics: 7.29% paved and 63.38% unpaved roads based on predictions.
• Augments OSM data with additional 0.0004 million km of surface information from Mapillary.

Limitations

• Approximately 29.33% of road surface information is missing in the underlying OSM data.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

OpenStreetMap (OSM), augmented with deep learning predictions from Mapillary imagery.

Collection Method

Surface classification derived using a hybrid deep learning approach, combined with GHSU and AFRICAPOLIS urban layers.

Time Range

Mean timestamp of predictions (DL_mean_timestamp) is provided per feature.

Freshness

Last updated 2026-03-02 17:08:45.642488; freshness should be verified.

Geography

Mongolia