Match and Locate: low-frequency monocular odometry based on deep feature matching
This addresses the need for affordable and simplified robotic systems by reducing reliance on high-frequency sensors, though it is incremental as it builds on existing deep feature matching techniques.
The paper tackles the problem of accurate and robust pose estimation for robotic systems by introducing a monocular odometry method that works with extremely low-frequency video (around one frame per second), achieving competitive results with approximately 3° orientation error and 2m translation error in a visual localization challenge.
Accurate and robust pose estimation plays a crucial role in many robotic systems. Popular algorithms for pose estimation typically rely on high-fidelity and high-frequency signals from various sensors. Inclusion of these sensors makes the system less affordable and much more complicated. In this work we introduce a novel approach for the robotic odometry which only requires a single camera and, importantly, can produce reliable estimates given even extremely low-frequency signal of around one frame per second. The approach is based on matching image features between the consecutive frames of the video stream using deep feature matching models. The resulting coarse estimate is then adjusted by a convolutional neural network, which is also responsible for estimating the scale of the transition, otherwise irretrievable using only the feature matching information. We evaluate the performance of the approach in the AISG-SLA Visual Localisation Challenge and find that while being computationally efficient and easy to implement our method shows competitive results with only around $3^{\circ}$ of orientation estimation error and $2m$ of translation estimation error taking the third place in the challenge.