AnchorFlow-VIO improves visual-inertial positioning with fixed templates
AnchorFlow-VIO: Fixed-Template Monocular Visual-Inertial Odometry on LaMAria
Robotics
Summary
Locating where a camera is in the world using only one camera and motion sensors is hard, especially in tricky conditions like low light or moving platforms. The authors developed AnchorFlow-VIO, a system that tracks features in images using fixed templates to improve stability. They tested this system extensively on a dataset called LaMAria, showing varying success depending on conditions and configurations. Their results highlight challenges and limits in current visual-inertial localization methods and provide open-source tools for further work.
What this means in practice
- •For robotics engineers: Enhance robotic navigation systems by integrating AnchorFlow-VIO to improve pose estimation from a single camera and inertial sensors under varied lighting and motion conditions.
- •For augmented reality developers: Implement improved visual-inertial tracking in AR devices using fixed-template methods for more reliable environment localization in challenging scenarios.
Authors
Sudhir Pratap Yadav
Abstract
We present AnchorFlow-VIO, an open-source Basalt-derived monocular visual-inertial odometry system, and evaluate a uniform configuration on LaMAria. The system uses one grayscale camera and one IMU, combining factory IMU correction, a fixed timing adjustment, windowed initialization and fixed-template patch tracking. Server evaluation of all 63 test recordings yields control-point scores of 47.2, 37.1, 24.8, 35.1 and 15.6 for Short, Medium, Long, Low light and Moving platform. Test pose recalls at 5 m are 84.6%, 69.1%, 53.9% and 62.9% for the four categories reporting dense recall. On ten training recordings, mean control-point score is 41.38 and all 265,694 image timestamps receive poses, including 204 filled poses. One training platform recording diverges and scores zero. A paired four-record comparison favors fixed over rolling templates on three recordings, while reversing that ranking on the fourth. Increasing feature density also yields mixed results. Ten fresh runs at prospectively fixed 30 s and 120 s offsets expose substantial start sensitivity: a low-light recording scores 6.14 at the earlier start and 42.00 at the later start. These results characterize the limits of the frozen system and show why complete output, feature counts and selected successful runs are insufficient evidence of robust localization. Source code, frozen configuration and reproducible submission tools are publicly available; training ablations are reported separately from the submitted test evaluation.