点云配准
点云配准是三维计算机视觉与摄影测量中的基本问题，旨在通过找到最佳变换矩阵，将不同坐标系下的多个点集对齐至同一坐标系。该技术在3D模型重建、文化遗产管理、滑坡监测及太阳能分析等众多视觉应用中具有重要价值。
ETH (trained on 3DMatch)
GeDi
3DMatch Benchmark
IMFNet
3DMatch (at least 30% overlapped - FCGF setting)
NgeNet
KITTI (trained on 3DMatch)
GeDi
3DLoMatch (10-30% overlap)
GeoTransformer - P2PNet
3DMatch (at least 30% overlapped - sample 5k interest points)
NgeNet
KITTI (FCGF setting)
FCGF
FPv1
Greedy Grid Search
FP-O-E
FP-O-H
FP-O-M
FP-R-E
FP-R-H
GeDi
FP-R-M
FP-T-E
FP-T-H
FP-T-M
KITTI
GeDi
RotKITTI Registration Benchmark
UMERegRobust
3DMatch (trained on KITTI)
GeDi
KITTI (Distant PCR)
nuScenes (Distant PCR)
GCL+KPConv
3RScan
SG-PGM