Northern Mariana Islands Road Surface and Passability AI Predictions (2020-2024)

HeiGIT produced this geospatial dataset covering approximately 100 km of arterial roads in the Northern Mariana Islands using 2020 and 2024 PlanetScope satellite imagery. It features AI-derived attributes for road surface, width, and passability to supplement missing OpenStreetMap (OSM) data. The analysis focuses specifically on motorway, trunk, primary, and secondary road classifications.

Use Cases

• Assessing logistical constraints for emergency response using the Humanitarian Passability Index (HPI) scores
• Identifying infrastructure development trends by comparing DL Surface Prediction 2020 and DL Surface Prediction 2024
• Planning heavy vehicle routes based on Road Width Class (1-3) to ensure unimpeded flow for trucks

Strengths

• High surface prediction accuracy of 89.2% compared to 64.7% for standard OSM tags
• Multi-temporal coverage (2020 and 2024) enables road surface change detection
• Standardized Humanitarian Passability Index (HPI) for logistical planning

Limitations

• Restricted to arterial road classes, excluding local and residential streets
• Small geographic footprint with only 100 km of total road length analyzed
• AI-derived predictions may contain errors in areas with dense canopy or cloud cover

Provenance

Source

Heidelberg Institute for Geoinformation Technology (HeiGIT)

Collection Method

Deep learning analysis of PlanetScope satellite imagery integrated with OpenStreetMap road network data

Time Range

2020-2024

Freshness

Contains data from 2020 and 2024; last updated March 2026.

Geography

Northern Mariana Islands