Topometric Localization with Deep Learning
This work addresses the need for affordable and reliable localization in applications like robotics or autonomous systems, offering a significant improvement over existing vision-based methods, though it is incremental as it builds on combining known techniques.
The paper tackles the problem of achieving high-accuracy localization using cost-effective cameras instead of expensive LiDAR sensors, by training deep networks on LiDAR-based data and combining visual odometry with topological localization, resulting in localization errors up to 10 times smaller than traditional vision-based methods.
Compared to LiDAR-based localization methods, which provide high accuracy but rely on expensive sensors, visual localization approaches only require a camera and thus are more cost-effective while their accuracy and reliability typically is inferior to LiDAR-based methods. In this work, we propose a vision-based localization approach that learns from LiDAR-based localization methods by using their output as training data, thus combining a cheap, passive sensor with an accuracy that is on-par with LiDAR-based localization. The approach consists of two deep networks trained on visual odometry and topological localization, respectively, and a successive optimization to combine the predictions of these two networks. We evaluate the approach on a new challenging pedestrian-based dataset captured over the course of six months in varying weather conditions with a high degree of noise. The experiments demonstrate that the localization errors are up to 10 times smaller than with traditional vision-based localization methods.