arXiv 2604.129422026-04-21cs.RO

RMGS-SLAM：基于3D高斯散射的实时多传感器SLAM

RMGS-SLAM: Real-time Multi-sensor Gaussian Splatting SLAM

Advanced Robotics Centre, National University of Singapore（新加坡国立大学先进机器人中心）；School of Aeronautics and Astronautics, Shanghai Jiao Tong University（上海交通大学航空宇航学院）；College of Computer Science, Sichuan University（四川大学计算机学院）；Singapore-MIT Alliance for Research and Technology（新加坡-麻省理工联合研究技术联盟）

Dongen Li, Yi Liu, Junqi Liu, Zewen Sun, Zefan Huang, Shuo Sun, Jiahui Liu, Chengran Yuan, Hongliang Guo, Francis E. H. Tay, Marcelo H. Ang

备注：: The manuscript has been improved, with refined content and updated and corrected experimental results

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出一种实时多传感器SLAM框架，结合激光雷达、惯性导航和视觉数据，通过高斯散射实现高精度定位与真实感映射，解决了大规模环境中的实时SLAM挑战。

摘要：

在大规模真实环境中实现基于3D高斯散射（3DGS）的实时同时定位与建图（SLAM）仍然具有挑战性，因为现有方法仍难以同时实现低延迟姿态估计、连续3D高斯重建和长期全局一致性。本文提出一种紧密耦合的激光雷达-惯性-视觉3DGS基于SLAM框架，用于实时姿态估计和真实感映射。系统通过并行执行状态估计和3D高斯原始初始化，与全局高斯优化协同工作，实现连续密集映射。为提高高斯初始化质量和加速优化收敛，我们引入了一种级联策略，结合前馈预测与基于体素的主成分分析导出的几何先验。为增强全局一致性，我们通过基于高斯的广义迭代最近点注册估计循环约束，直接在优化后的全局高斯地图上执行循环闭合，随后进行位姿图优化。我们还收集了具有硬件同步的激光雷达-摄像头-IMU和真实轨迹的挑战性大规模循环户外序列进行评估。在公共数据集和我们自己的数据集上的广泛实验表明，所提方法在实时效率、定位精度和渲染质量方面在多种真实场景中达到最佳状态。

英文摘要：

Achieving real-time Simultaneous Localization and Mapping (SLAM) based on 3D Gaussian splatting (3DGS) in large-scale real-world environments remains challenging, as existing methods still struggle to jointly achieve low-latency pose estimation, continuous 3D Gaussian reconstruction, and long-term global consistency. In this paper, we present a tightly coupled LiDAR-Inertial-Visual 3DGS-based SLAM framework for real-time pose estimation and photorealistic mapping in large-scale real-world scenes. The system executes state estimation and 3D Gaussian primitive initialization in parallel with global Gaussian optimization, enabling continuous dense mapping. To improve Gaussian initialization quality and accelerate optimization convergence, we introduce a cascaded strategy that combines feed-forward predictions with geometric priors derived from voxel-based principal component analysis. To enhance global consistency, we perform loop closure directly on the optimized global Gaussian map by estimating loop constraints through Gaussian-based Generalized Iterative Closest Point registration, followed by pose-graph optimization. We also collect challenging large-scale looped outdoor sequences with hardware-synchronized LiDAR-camera-IMU and ground-truth trajectories for realistic evaluation. Extensive experiments on both public datasets and our dataset demonstrate that the proposed method achieves a state of the art among real-time efficiency, localization accuracy, and rendering quality across diverse real-world scenes.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。