Deep Feature-Based Visual-Inertial Odometry

Introduction

We present a hybrid visual-inertial odometry system that relies on a state-of-the-art deep feature matching front-end and a traditional visual-inertial optimization back-end. More precisely, we develop a fully-fledged feature tracker based on the recent SuperPoint and LightGlue neural networks, that can be plugged directly to the estimation back-end of VINS-Mono. By default, this feature tracker returns extremely abundant matches. To bound the computational complexity of the back-end optimization, limiting the number of used matches is desirable. Therefore, we explore various methods to filter the matches while maintaining a high visual-inertial odometry performance. We run extensive tests on the EuRoC machine hall and Vicon room datasets, showing that our system achieves state-of-the-art odometry performance according to relative pose errors.

Links

• Repository: https://github.com/charleshamesse/LightGlue-VINS-Mono
• Citations:
  • Practical Deep Feature-Based Visual-Inertial Odometry
    C Hamesse, M Vlaminck, H Luong, R Haelterman
    13th International Conference on Pattern Recognition Applications and Methods (ICPRAM), 2024