Location Prior Generation via Multi-Source Urban Data Fusion for Low-Altitude Air Mobility
Researchers have developed a framework called LPGF to generate 3D urban spatial data, specifically building heights, which are missing from most global geospatial databases. This framework fuses data from sources like satellite imagery, UAV telemetry, GPS trajectories, and OpenStreetMap to create structured location priors. LPGF prioritizes explicit height tags, then floor counts, and finally default building-type heights, with an optional shadow-based estimation module for improved accuracy. AI
IMPACT This framework could enable more autonomous aerial operations in urban environments by providing crucial 3D spatial data.