ContrastMotion: Self-supervised Scene Motion Learning for Large-Scale LiDAR Point Clouds

In this paper, we propose a novel self-supervised motion estimator for LiDAR-based autonomous driving via BEV representation. Different from usually adopted self-supervised strategies for data-level structure consistency, we predict scene motion via feature-level consistency between pillars in consecutive frames, which can eliminate the effect caused by noise points and view-changing point clouds in dynamic scenes. Specifically, we propose $\mathit{\text{Soft}}\mathit{\text{Discriminative}}\mathit{\text{Loss}}$ that provides the network with more pseudo-supervised signals to learn discriminative and robust features in a contrastive learning manner. We also propose $\mathit{\text{Gated}}\mathit{\text{Multi-frame}}\mathit{\text{Fusion}}$ block that learns valid compensation between point cloud frames automatically to enhance feature extraction. Finally, $\mathit{\text{pillar}}\mathit{\text{association}}$ is proposed to predict pillar correspondence probabilities based on feature distance, and whereby further predicts scene motion. Extensive experiments show the effectiveness and superiority of our $\mathbf{\text{ContrastMotion}}$ on both scene flow and motion prediction tasks. The code is available soon.