UAVs locate themselves using building shapes in cities without GPS
Large-Scale Geometric Map-Based Localization of UAVs in GNSS-Denied Urban Environments
Robotics
Summary
Flying drones often rely on GPS to know where they are, but GPS can stop working well in cities surrounded by tall buildings. This paper presents a way for drones to figure out their location by looking at the shapes and arrangements of buildings from above. Instead of relying on how buildings look in photos, the method uses the geometric layout of buildings, which stays consistent even if the images change. The authors tested their system over large city areas and found it was highly accurate, outperforming other existing methods.
What this means in practice
- •For drone operators: Enable drones to reliably locate themselves in GPS-blocked urban areas by matching building layouts to a reference database.
- •For emergency response teams: Use drones to navigate and position themselves during operations in cities where GPS signals are unavailable or unreliable.
Authors
Garth Terlizzi, Kaveh Fathian
Abstract
Unmanned aerial vehicles (UAVs) operating in GNSS-denied urban environments require alternative methods for position estimation. Existing approaches based on satellite image retrieval or learned descriptors are sensitive to appearance variation and degrade rapidly as the search area grows. We present a vision-based localization system that matches building patterns observed from a downward-facing UAV camera against a reference building footprint database. Our approach detects buildings in aerial imagery, accumulates observations across frames into a unified map, and matches local building arrangements against reference footprints using a novel geometry-driven descriptor that augments local triangle structure with per-building shape features. By encoding spatial relationships between nearby buildings rather than visual appearance, the system is robust to appearance variations and remains discriminative over large search areas. Evaluations on seven flights across four municipalities in a large metropolitan area demonstrate 100% Recall@1 at search areas of approximately 113 km$^2$ and 254 km$^2$, and 71.4% Recall@1 when expanded to approximately 452 km$^2$, encompassing up to 277,000 buildings. In contrast, baseline methods degrade rapidly and achieve 0% Recall@1 at 254 km$^2$ and 452 km$^2$.