Towards End-to-end Learning of Monocular Visual-inertial Odometry with an Extended Kalman Filter
Author | : Chunshang Li |
Publisher | : |
Total Pages | : 0 |
Release | : 2020 |
ISBN-10 | : OCLC:1335043610 |
ISBN-13 | : |
Rating | : 4/5 (10 Downloads) |
Book excerpt: Classical visual-inertial fusion relies heavily on manually crafted image processing pipelines, which are prone to failure in situations with rapid motion and texture-less scenes. While end-to-end learning methods show promising results in addressing these limitations, embedding domain knowledge in the form of classical estimation processes within the end-to-end learning architecture has the potential of combining the best of both worlds. In this thesis, we propose the first end-to-end trainable visual-inertial odometry (VIO) algorithm that leverages a robo-centric Extended Kalman Filter (EKF). The EKF propagates states through a known inertial measurement unit (IMU) kinematics model and accepts relative pose measurements and uncertainties from a deep network as updates. The system is fully differentiable and can be trained end-to-end through backpropagation. Our method achieves competitive results among state of the art classical and learning based VIO methods on the KITTI dataset.