Kenya Arterial Roads: AI-Derived Surface and Passability for 21,100 km

Mapping 21,100 km of arterial roads in Kenya, this dataset provides AI-derived surface types, widths, and passability scores generated by HeiGIT. It integrates 2020 and 2024 PlanetScope satellite imagery with OpenStreetMap attributes to classify motorway, trunk, primary, and secondary road segments.

Use Cases

• Evaluating logistics routes for heavy trucks using 'Road Width Class' and 'Humanitarian Passability Index'
• Identifying infrastructure improvements by filtering for 'Road Surface Change' transitions
• Supplementing missing 'surface tag' data in OpenStreetMap with 'DL Surface Prediction'

Strengths

• 89.2% accuracy for surface model predictions
• Covers 21,100 km of arterial road infrastructure
• Multi-temporal comparison between 2020 and 2024 imagery

Limitations

• Excludes tertiary and residential road networks
• AI-derived predictions may have localized errors compared to ground-level surveys

Provenance

Source

HeiGIT (Heidelberg Institute for Geoinformation Technology)

Collection Method

Deep learning analysis of PlanetScope satellite imagery combined with OpenStreetMap vector data

Time Range

2020-2024

Freshness

Last updated March 2026; contains imagery-derived data from 2020 and 2024.

Geography

Kenya