GFT‐VINS: Robust Visual–Inertial Localization via Geometric Feature Track Selection
Journal of Field Robotics · 2026-09 · 43 (6) · 3979-3993
DOI: 10.1002/rob.70227 ↗At a glance
Journal article in Journal of Field Robotics, published online on 2026-04-27. Written by Hui Zhao and 7 co-authors. 48 references. Funded by: National Natural Science Foundation of China.
Abstract (original language)
Visual–inertial navigation systems (VINSs) are a cornerstone technology for autonomous robotics, leveraging visual measurements and inertial sensor data to achieve accurate state estimation through feature tracking. However, current VINS approaches often indiscriminately incorporate all feature tracks without distinguishing which ones meaningfully contribute to estimation quality. This oversight leads to suboptimal accuracy, reduced stability, and computational inefficiency. To address these limitations, we propose a novel visual–inertial framework that intelligently prioritizes informative feature tracks using normal epipolar geometry. Our method enhances state estimation by jointly optimizing feature appearance similarity and observation consistency across frames, formalized as a submodular partition optimization problem. Extensive experiments on public benchmarks demonstrate that our approach significantly outperforms five state‐of‐the‐art methods, achieving a 26.1% improvement in accuracy, 43.3% higher stability, and real‐time processing speeds of up to 125 FPS. These advancements highlight the efficacy of selective feature track utilization in overcoming the limitations of conventional VINS.
Funding
- National Natural Science Foundation of China
Cite
APA
Zhao, H., Dou, J., Shang, J., Ji, K., Li, Y., Li, Y., Khoshelham, K., & Gu, F. (2026). GFT‐VINS: Robust Visual–Inertial Localization via Geometric Feature Track Selection. Journal of Field Robotics, 43(6), 3979-3993. https://doi.org/10.1002/rob.70227
Vancouver
Zhao H, Dou J, Shang J, Ji K, Li Y, Li Y, et al. GFT‐VINS: Robust Visual–Inertial Localization via Geometric Feature Track Selection. Journal of Field Robotics. 2026;43(6):3979-3993. doi:10.1002/rob.70227
BibTeX
@article{Zhao2026,
author = {Zhao, Hui and Dou, Jiarui and Shang, Jianga and Ji, Kangping and Li, You and Li, Yan and Khoshelham, Kourosh and Gu, Fuqiang},
title = {GFT‐VINS: Robust Visual–Inertial Localization via Geometric Feature Track Selection},
journal = {Journal of Field Robotics},
year = {2026},
volume = {43},
number = {6},
pages = {3979--3993},
doi = {10.1002/rob.70227}
}
EUVIMED