DeepMapping2: Self-Supervised Large-Scale LiDAR Map Optimization
This work addresses the problem of accurate global point cloud registration for self-driving and mobile robotics, representing an incremental improvement over prior methods.
The paper tackles the challenge of large-scale LiDAR map optimization by proposing DeepMapping2, which improves upon DeepMapping by incorporating loop closures and point consistency losses, achieving better performance on datasets like KITTI, NCLT, and Nebula.
LiDAR mapping is important yet challenging in self-driving and mobile robotics. To tackle such a global point cloud registration problem, DeepMapping converts the complex map estimation into a self-supervised training of simple deep networks. Despite its broad convergence range on small datasets, DeepMapping still cannot produce satisfactory results on large-scale datasets with thousands of frames. This is due to the lack of loop closures and exact cross-frame point correspondences, and the slow convergence of its global localization network. We propose DeepMapping2 by adding two novel techniques to address these issues: (1) organization of training batch based on map topology from loop closing, and (2) self-supervised local-to-global point consistency loss leveraging pairwise registration. Our experiments and ablation studies on public datasets (KITTI, NCLT, and Nebula) demonstrate the effectiveness of our method.