Neural LiDAR Bundle Adjustment

2026-07-05

AI Summary

This work addresses the critical influence of volumetric sampling density on multi-view LiDAR point cloud registration and 3D mapping performance within LiDAR-based Neural Radiance Fields (LiDAR NeRF). To this end, we propose an efficient voxel sampling strategy tailored to LiDAR rays and introduce NeLD-BA, a neural bundle adjustment algorithm specifically designed for the LiDAR modality. Our method jointly optimizes sensor poses and the neural radiance field map, substantially improving both registration accuracy and reconstruction quality. Experimental results on the Newer College and FusionPortable datasets demonstrate that the proposed approach achieves state-of-the-art performance in multi-view LiDAR point cloud registration and 3D reconstruction tasks.

Abstract

Recent research has achieved remarkable novel view rendering and scene reconstruction results with Neural Radiance Field (NeRF), including extensions to the LiDAR modality. Few studies have, however, explored the key design differences between RGB NeRFs and LiDAR NeRFs, particularly considering their underlying working principles. In this work, we provide both theoretical and empirical evidence suggesting that the density of volume sampling plays a significant role in LiDAR NeRF. Based on this finding, we propose a novel Neural LiDAR Bundle Adjustment (NeLD-BA) algorithm, which is tailored using efficient volume sampling of LiDAR rays for joint optimization of LiDAR map and poses. Extensive experiments are performed using the Newer College and FusionPortable datasets to demonstrate the proposed NeLD-BA's state-of-the-art performance in multi-view point cloud registration and 3D mapping. We will open-source our code for the community.

Problem

• Neural Radiance Field
• LiDAR
• Bundle Adjustment
• 3D Mapping
• Point Cloud Registration

Innovation

• Neural LiDAR Bundle Adjustment
• volume sampling
• LiDAR NeRF
• joint optimization
• 3D mapping