IRAL: Robust and versatile UAV localization using infrared vision and altitude sensor fusion

Yixian Li, Qiang Wang*, Zhonghu Hao, Shengrong Hu, Jiaxing Wu, Linkang Dong

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Supplementing or replacing the Global Satellite Navigation System (GNSS) for robust UAV localization remains a challenge. In this work, we propose an infrared vision and altitude sensor fusion method called IRAL, which mainly includes on-board near-infrared (NIR) beacons, off-board narrowband-pass vision sensors with the same wavelength as the beacon and strapdown high precision altitude sensors. The beacon with a high signal-to-noise ratio as a cooperative target provides robust features, thereby facilitating beacon recognition through the designed gradient-based sequential frame template matching (GSFTM) algorithm. The proposed method measures the altitude difference between the UAV and the vision sensor through the altitude sensor to accomplish depth estimation. After obtaining the beacon's pixel coordinates and depth, combined with the intrinsics and extrinsics of the vision sensor, the observation equation can be set up to solve the UAV's spatial position. Real-world experiments under various scenarios demonstrate that the proposed method stably achieves high accuracy.

Original languageEnglish
Article number115917
JournalMeasurement: Journal of the International Measurement Confederation
Volume242
DOIs
Publication statusPublished - Jan 2025

Keywords

  • Cooperative method
  • Infrared vision
  • Localization
  • Sensor fusion
  • Unmanned aerial vehicle

Fingerprint

Dive into the research topics of 'IRAL: Robust and versatile UAV localization using infrared vision and altitude sensor fusion'. Together they form a unique fingerprint.

Cite this