arXiv 2505.099152025-05-15cs.CVcs.RO

大规模 Gaussian Splatting SLAM

Large-Scale Gaussian Splatting SLAM

Meituan UAV（美团无人机）；School of Computer Science and Technology, University of Science and Technology of China（中国科学技术大学计算机科学与技术学院）；Dept. of Mechanical Engineering, Computer and Information Sciences, University of Delaware（德雷塞尔大学机械工程、计算机与信息科学系）

Zhe Xin, Chenyang Wu, Penghui Huang, Yanyong Zhang, Yinian Mao, Guoquan Huang

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出面向大规模户外场景的 LSG-SLAM，融合多模态位姿估计、特征对齐跟踪、连续 GS 子地图与回环优化，在 EuRoc 和 KITTI 上优于多种现有方法。

摘要：

近期发展的 Neural Radiance Fields（NeRF，神经辐射场）和 3D Gaussian Splatting（3DGS，三维高斯泼溅）已在视觉 SLAM 中展现出令人鼓舞且印象深刻的结果。然而，大多数代表性方法都需要 RGBD 传感器，并且仅适用于室内环境。大规模户外场景中重建的鲁棒性仍未得到探索。本文提出一种基于 3DGS、使用双目相机的大规模视觉 SLAM，称为 LSG-SLAM。所提出的 LSG-SLAM 采用多模态策略，在较大视角变化下估计先验位姿。在跟踪中，我们引入特征对齐变形约束，以缓解渲染损失中外观相似性带来的不利影响。为了实现大规模场景的可扩展性，我们引入连续的 Gaussian Splatting 子地图，以有限内存处理无界场景。通过场景识别在 GS 子地图之间检测回环，并利用渲染和特征变形损失优化回环关键帧之间的相对位姿。在相机位姿和高斯点进行全局优化后，结构精修模块提升重建质量。通过在 EuRoc 和 KITTI 数据集上的广泛评估，LSG-SLAM 相比现有基于 Neural、基于 3DGS，甚至传统方法都取得了更优性能。项目主页：https://lsg-slam.github.io。

英文摘要：

The recently developed Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have shown encouraging and impressive results for visual SLAM. However, most representative methods require RGBD sensors and are only available for indoor environments. The robustness of reconstruction in large-scale outdoor scenarios remains unexplored. This paper introduces a large-scale 3DGS-based visual SLAM with stereo cameras, termed LSG-SLAM. The proposed LSG-SLAM employs a multi-modality strategy to estimate prior poses under large view changes. In tracking, we introduce feature-alignment warping constraints to alleviate the adverse effects of appearance similarity in rendering losses. For the scalability of large-scale scenarios, we introduce continuous Gaussian Splatting submaps to tackle unbounded scenes with limited memory. Loops are detected between GS submaps by place recognition and the relative pose between looped keyframes is optimized utilizing rendering and feature warping losses. After the global optimization of camera poses and Gaussian points, a structure refinement module enhances the reconstruction quality. With extensive evaluations on the EuRoc and KITTI datasets, LSG-SLAM achieves superior performance over existing Neural, 3DGS-based, and even traditional approaches. Project page: https://lsg-slam.github.io.

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。