Efficient and robust point cloud matching in low-overlap scene scans
Spectral Consistency-Guided Multiview Point Cloud Registration for Low-Overlap Scenes
Computer Vision and Pattern Recognition
Summary
Matching 3D scans from different views can be hard when they barely overlap because it's tough to find common points. The authors developed a method called GMPCR that picks the most trustworthy matches and scan pairs using a special math technique involving spectra. This helps reduce errors and speeds up matching by ignoring less useful scan pairs. Their approach works well on various real-world datasets, improving the accuracy and efficiency of aligning multiple 3D scans.
What this means in practice
- •For autonomous vehicle developers: Improve the alignment of 3D lidar scans from different viewpoints in environments with limited overlap, enhancing vehicle perception and mapping accuracy.
- •For robotics system integrators: Create faster and more reliable composite maps from multiple partial 3D scans in environments where scan overlap is minimal or uncertain.
Authors
Tianyu Li, Yanghong Lin, Shudong Zhou, Kui Yang, Jingru Zhang, Li Fang, Wei Yao
Abstract
Multiview point cloud registration is particularly challenging in low-overlap scenes, where reliable correspondences are limited and incorrect pairwise transformations can affect global pose estimation. In addition, registering all scan pairs is computationally expensive because many pairs provide weak geometric information. To address these problems, we propose GMPCR, a non-learning-based spectral consistency-guided framework for efficient and robust multiview point cloud registration. GMPCR builds a refined second-order compatibility structure from initial correspondences and uses its dominant spectral response to evaluate both correspondence reliability and scan-pair confidence. This allows unreliable correspondences to be filtered and informative scan pairs to be selected before relative transformation estimation, leading to a sparse pose graph and reduced pairwise registration cost. For each retained scan pair, maximal-clique-based hypothesis generation is used to estimate reliable relative transformations. The resulting pose graph is further refined by an adaptive history-aware synchronization scheme, in which the effect of residual history is adjusted according to changes in the global rotation residual. A recovery mechanism also allows down-weighted edges to regain confidence when their global consistency improves. Experiments on 3DMatch, 3DLoMatch, ScanNet, and ETH demonstrate the effectiveness of GMPCR. It achieves registration recalls of 97.2% and 89.6% on 3DMatch and 3DLoMatch, respectively, while maintaining competitive performance on ScanNet and ETH. The results show that GMPCR provides a favorable balance among registration accuracy, robustness to low overlap, and computational efficiency. The code is publicly available at https://github.com/swccj/gmpcr.