Contrastive learning-based absolute visual localization for UAVs
Loading...
Date
2026-04-28
Authors
Κωνσταντακόπουλος, Αθανάσιος
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Unmanned Aerial Vehicles (UAVs) typically rely on external signals, such as GNSS signals, to accurately calculate their geographical position. However, in situations where such signals are unreliable or entirely unavailable, visual localization becomes crucial. It enables the UAV to estimate either its own position or the position of a target, using data captured by an onboard camera, by matching or registering UAV-view images against a geo-referenced database of satellite imagery. In the context of this thesis, the research area of Absolute Visual Localization (AVL) is analyzed in detail, with a focus on the most prominent contrastive learning-based methodologies.
Description
Keywords
UAVs, Visual localization, Deep learning