Geometry-Preserving in 3D Gaussian Splatting for LiDAR-Camera Extrinsic Calibration

2026-06-18

AI Summary

This work addresses the challenge of insufficient accuracy in extrinsic calibration between LiDAR and cameras in markerless scenarios, primarily caused by sparse cross-modal correspondences. To overcome this limitation, we propose a geometry-faithful joint optimization method based on 3D Gaussian splatting. By incorporating multi-view LiDAR depth supervision and freezing photometric gradients during differentiable rendering to prevent updates to Gaussian spatial parameters, our approach ensures that the proxy geometry remains consistent with the true LiDAR structure, thereby avoiding interference from rendering-based optimization on calibration accuracy. The method enables end-to-end extrinsic calibration and demonstrates consistently superior performance over existing markerless approaches on public autonomous driving datasets, achieving significant and stable improvements in calibration precision.

Abstract

Accurate LiDAR-camera calibration is essential for robust multi-modal perception. Targetless approaches avoid manual setup but remain limited by the scarcity of discriminative cross-modal features. Recent methods address this by reconstructing the scene within a differentiable model, enabling extrinsic optimization through dense photometric supervision. Among these, 3D Gaussian Splatting (3DGS) has been widely adopted as a geometric proxy that bridges LiDAR and camera within a single differentiable framework. However, since 3DGS was originally designed for novel view synthesis, existing methods tend to prioritize rendering quality, causing the proxy geometry to drift from the true LiDAR structure. We propose a framework that preserves the metric geometry of the Gaussian proxy by aggregating multi-view LiDAR observations for dense depth supervision and blocking photometric gradients from updating the Gaussian spatial parameters. We validate our method on public driving datasets, where it consistently outperforms existing targetless methods in calibration accuracy.

Problem

• LiDAR-camera calibration
• 3D Gaussian Splatting
• geometry preservation
• targetless calibration
• extrinsic calibration

Innovation

• geometry-preserving
• 3D Gaussian Splatting
• LiDAR-camera calibration
• targetless calibration
• dense depth supervision