GeoGS-SLAM: Geometry-Only Gaussian Splatting for Dense Monocular SLAM

📄 arXiv: 2607.07452v1 📥 PDF

作者: Lipu Zhou, Yaoyun Kang, Junxiang Pang, Shengkai Sun, Tingting Bao, Kehan Wang

分类: cs.RO

发布日期: 2026-07-08


💡 一句话要点

提出GeoGS-SLAM以解决稠密单目SLAM中的几何重建问题

🎯 匹配领域: 支柱三:空间感知与语义 (Perception & Semantics)

关键词: 稠密视觉SLAM 几何重建 高斯点云 机器人导航 光照鲁棒性 在线映射 训练框架

📋 核心要点

  1. 现有的3DGS方法在处理稠密SLAM时,往往同时关注外观和几何建模,导致效率低下。
  2. 我们提出的GeoGS方法专注于几何重建,显著减少了高斯原始体的数量,提高了几何收敛速度。
  3. 在合成和真实世界基准测试中,我们的方法在在线映射效率和几何重建质量上均优于现有最先进技术。

📝 摘要(中文)

稠密视觉SLAM是机器人领域的一个基本问题。近年来,3DGS的进展展示了其在稠密SLAM中的潜力。现有的3DGS框架关注于外观和几何建模,但在SLAM中,场景几何通常比新视图合成更为重要。基于此,我们提出了仅基于几何的高斯点云(GeoGS),并进一步提出了基于该表示的稠密视觉SLAM系统GeoGS-SLAM。GeoGS仅保留空间参数,减少了80%以上的每个原始参数数量,专注于几何重建,显著加快了几何收敛,并增强了对光照变化的鲁棒性。我们还提出了一种有效的训练框架,通过单视图和多视图的几何和光度监督来优化高斯原始体,并通过局部平面驱动的初始化加速几何收敛。实验表明,我们的方法在在线映射效率和几何重建质量上超越了现有最先进的方法。

🔬 方法详解

问题定义:本论文旨在解决稠密单目SLAM中的几何重建问题,现有方法在处理场景几何时往往过于依赖外观建模,导致效率低下和鲁棒性不足。

核心思路:我们提出的GeoGS方法仅关注几何重建,去除了对场景外观的建模,从而减少了高斯原始体的数量,提升了几何收敛速度和鲁棒性。

技术框架:GeoGS-SLAM系统由几个主要模块组成,包括几何重建模块、训练框架和地图更新策略。训练框架结合单视图和多视图的几何与光度监督,优化高斯原始体。

关键创新:GeoGS的核心创新在于其仅保留空间参数,减少了80%以上的每个原始参数数量,专注于几何重建,显著提升了系统的效率和鲁棒性。

关键设计:我们设计了局部平面驱动的初始化方法,以更好地对齐原始体与局部结构,同时采用了新的损失函数来优化高斯原始体的几何和光度特性。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

在实验中,GeoGS-SLAM在合成和真实世界基准测试中表现出色,在线映射效率和几何重建质量均优于现有最先进的方法,具体性能数据表明,几何收敛速度提高了显著的百分比,且在光照变化下的鲁棒性得到了增强。

🎯 应用场景

该研究的潜在应用领域包括机器人导航、障碍物避让和增强现实等。通过提高稠密SLAM的效率和几何重建质量,GeoGS-SLAM能够在复杂环境中实现更高效的实时定位与地图构建,具有重要的实际价值和未来影响。

📄 摘要(原文)

Dense visual SLAM is a fundamental problem in robotics. Recent advances in 3DGS have demonstrated its potential for dense SLAM. Existing 3DGS frameworks focus on both appearance and geometry modeling. However, scene geometry is typically more critical for SLAM than novel view synthesis because downstream robotic tasks, such as navigation and obstacle avoidance, rely primarily on accurate spatial geometry rather than photorealistic rendering. This observation raises a natural question: Is it feasible for 3DGS to perform 3D reconstruction without scene appearance modeling? Motivated by this, we propose Geometry-only Gaussian Splatting (GeoGS), which directly reconstructs scene geometry, and further present GeoGS-SLAM, a dense visual SLAM system built upon this representation. Specifically, GeoGS retains only spatial parameters to reduce the number of per-primitive parameters by over 80%. In contrast to existing 3DGS methods, GeoGS focuses solely on geometric reconstruction, which significantly reduces the number of Gaussian primitives, accelerates geometric convergence, and enhances robustness to illumination variations. In addition, we present an effective training framework that optimizes the Gaussian primitives via single-view and multi-view geometric and photometric supervision, and speeds up geometry convergence with a local-plane driven initialization that better aligns primitives with local structures. Furthermore, we introduce a map update strategy for loop closure that globally transforms the Gaussian map to align it with the corrected pose estimates, thereby preventing map tearing caused by inconsistent per-viewpoint pose corrections in existing methods. Extensive experiments on synthetic and real-world benchmarks demonstrate that our method outperforms SOTA methods in terms of online mapping efficiency and geometric reconstruction quality.