Paired-CSLiDAR: Height-Stratified Registration for Cross-Source Aerial-Ground LiDAR Pose Refinement

2026-05-01

AI Summary

This study addresses the challenge of pose optimization in single-scan alignment between airborne and terrestrial LiDAR point clouds, which often suffers from limited geometric overlap and convergence to local optima. To facilitate fine-grained registration within a 50-meter range, the authors construct a large-scale benchmark dataset comprising 12,683 scan pairs. They propose a training-free, geometry-driven optimization pipeline that robustly estimates poses by focusing on shared terrain surfaces. The method innovatively integrates elevation-stratified ICP, a bidirectional registration mechanism, residual-guided refinement, and a confidence-gated accept-or-better selection strategy. Evaluated on 9,012 test samples, the approach achieves success rates of 86.0% at a 0.75-meter threshold (S@0.75m) and 99.8% at 1.0 meter (S@1.0m), significantly outperforming existing methods such as GeoTransformer.

Abstract

We introduce Paired-CSLiDAR (CSLiDAR), a cross-source aerial-ground LiDAR benchmark for single-scan pose refinement: refining a ground-scan pose within a 50 m-radius aerial crop. The benchmark contains 12,683 ground-aerial pairs across 6 evaluation sites and per-scan reference 6-DoF alignments for sub-meter root-mean-square error (RMSE) evaluation. Because aerial scans capture rooftops and canopy while ground scans capture facades and under-canopy, the two modalities share only a fraction of their geometry, primarily the terrain surface, causing standard registration methods and learned correspondence models to converge to metrically incorrect local minima. We propose Residual-Guided Stratified Registration (RGSR), a training-free, geometry-only refinement pipeline that exploits the shared ground plane through height-stratified ICP, reversed registration directions, and confidence-gated accept-if-better selection. RGSR achieves 86.0% S@0.75 m and 99.8% S@1.0 m on the primary benchmark of 9,012 scans, outperforming both the confidence-gated cascade at 83.7% and GeoTransformer at 76.3%. We validate RMSE-based pose selection with independent survey control and trajectory consistency, and show that added Fourier-Mellin BEV proposals can reduce RMSE while increasing actual pose error under extreme partial overlap. The dataset and code are being prepared for public release.

Problem

• cross-source LiDAR
• pose refinement
• aerial-ground registration
• partial overlap
• local minima

Innovation

• height-stratified registration
• cross-source LiDAR
• pose refinement
• geometry-only registration
• residual-guided ICP