Indonesia Road Surface Classification from OpenStreetMap and AI

Approximately 1.8582 million kilometers of roads in Indonesia are mapped, with AI-derived surface classifications for paved and unpaved segments. This dataset, created by the Heidelberg Institute for Geoinformation Technology (HeiGIT), combines OpenStreetMap data with deep learning predictions from Mapillary imagery and urban layers. It was last updated in March 2026.

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 based on the proportion of paved versus unpaved roads.
• GIS applications leveraging the geospatial road network and urban/rural classification.

Strengths

• Covers approximately 1.8582 million kilometers of roads in Indonesia.
• Provides AI-derived surface classifications for 9.5052% paved and 6.877% unpaved roads.
• Integrates multiple data sources: OpenStreetMap, Mapillary imagery, and urban settlement layers.

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.
• About 83.6178% of road surface information is reported as 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 included in the 'DL_mean_timestamp' field.

Freshness

Last updated 2026-03-02 17:16:06.105869; freshness should be verified.

Geography

Indonesia