A Registration-aided Domain Adaptation Network for 3D Point Cloud Based Place Recognition

TL;DR

This paper introduces a registration-aided domain adaptation network for 3D point cloud place recognition, leveraging synthetic data and geometric registration to improve real-world performance in autonomous navigation.

Contribution

It proposes a novel registration-aided domain adaptation network that uses synthetic data and geometric registration to enhance 3D place recognition in real-world environments.

Findings

Outperforms state-of-the-art baselines on Oxford RobotCar dataset.

Achieves comparable results to existing methods with registration visualization.

Effectively uses synthetic GTA-V data for training.

Abstract

In the field of large-scale SLAM for autonomous driving and mobile robotics, 3D point cloud based place recognition has aroused significant research interest due to its robustness to changing environments with drastic daytime and weather variance. However, it is time-consuming and effort-costly to obtain high-quality point cloud data for place recognition model training and ground truth for registration in the real world. To this end, a novel registration-aided 3D domain adaptation network for point cloud based place recognition is proposed. A structure-aware registration network is introduced to help to learn features with geometric information and a 6-DoFs pose between two point clouds with partial overlap can be estimated. The model is trained through a synthetic virtual LiDAR dataset through GTA-V with diverse weather and daytime conditions and domain adaptation is implemented to the real-world domain by aligning the global features. Our results outperform state-of-the-art 3D place recognition baselines or achieve comparable on the real-world Oxford RobotCar dataset with the visualization of registration on the virtual dataset.