Conference
Vision-Based Unmanned Aerial Vehicle Pose Correction Using a Digital Model
2026 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC)
— 2026 — IEEE
Key information
Authors:
Published in
May 20, 2026
Abstract
Unmanned Aerial Vehicles (UAVs) are extensively used in remote sensing applications such as wildfire monitoring, where precise georeferencing is critical. However, onboard navigation sensors, including Global Navigation Satellite System (GNSS) receivers and Inertial Measurement Units (IMUs), are prone to errors that can affect downstream processing. This paper introduces a UAV pose correction framework based on synthetic view generation and image-based matching. UAV telemetry, smoothed with a constant-velocity Kalman filter, is used to generate synthetic views, which are then matched to real UAV images using Detector-Free Local Feature Matching with Transformers (LoFTR). The correspondences obtained are ray-cast into a digital surface model to recover 3D points, which are subsequently employed for camera pose refinement via Perspective-n-Point estimation using the Sequential Quadratic Programming PnP (SQPnP) algorithm. The approach is tested on an aerial dataset collected over a forested area at an average altitude of 1678 m. Experimental results reveal horizontal and vertical RMSEs of 3.32 m and 2.21 m, respectively, with orientation errors below 0.2°. Further refinement using an Improved Innovation Adaptive Kalman Filter (IAKF) reduces the horizontal and vertical RMSEs to 2.24 m and 1.01 m, respectively. These findings demonstrate the effectiveness of the proposed method for robust UAV pose correction.
Publication details
Authors in the community:
Publication version
AM - Accepted manuscript
Publisher
IEEE
Link to the publisher's version
https://ieeexplore.ieee.org/abstract/document/11523313
Title of the publication container
2026 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC)
Location of the conference
Barcelos, Portugal
Conference date start
April 22, 2026
Conference date end
April 23, 2026
Fields of Science and Technology (FOS)
electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering
Keywords
- Autonomous aerial vehicles
- Modeling
- Filtering
- Telemetry
- Cameras
- Kalman filters
- Digital Model
- Pose Correction
- Camera pose estimation
- Pose estimation
- Temporal Filtering
Publication language (ISO code)
eng - English
Rights type:
Only metadata available
Financing entity
Fundação para a Ciência e a Tecnologia
Title of the project, award or grant: LA/P/0083/2020, UIDP/50009/2020, UIDB/50009/2020
Identifier for the funding entity: https://doi.org/10.13039/501100001871
Type of identifier of the funding entity: Crossref Funder