Developed by UN Global Pulse and UNOSAT, PulseSatellite is a collaborative toolkit using human-AI feedback loops to analyze satellite imagery quickly across disasters—mapping shelters, floods, and damage in crisis zones.
Jan 12, 2020
Problem & Need
In disasters, responders must quickly interpret satellite data to map displacement, flood zones, and infrastructure damage. Traditional manual analysis is slow and resource-intensive.
What PulseSatellite Does
It loads neural network models (structure mapping, flood segmentation, roof density) and allows human users to correct / retrain the model on the fly. unglobalpulse.org+1Impact / Use Cases
• Mapping refugee shelters / settlements
• Flood mapping in disaster zones
• Roof / building density detection for neighborhood structure analysis unglobalpulse.org+1Benefits / Innovations
• Human-in-the-loop makes the system adapt to local terrain and context. Medium+2unglobalpulse.org+2
• Speeds up analysis time and reduces manual burden for UN analysis teams. unglobalpulse.org+1Lessons / Takeaways for AstroSentinel
• Blending AI + human correction is powerful — perfect for uncertain terrain / species detection.
• Need modular models (you’ll want to support displacement, habitat, species)
• A system that adapts / retrains in context is more resilient.
Sign up to hear about pilots, partnerships, and early access opportunities.