Robust and Efficient Monocular 3D Gaussian SLAM for Kilometer-Scale Outdoor Scenes

2026-06-29

AI Summary

This work addresses the challenges of fragile long-term tracking and excessive memory consumption in monocular 3D Gaussian SLAM for kilometer-scale outdoor scenes by introducing KiloGS-SLAM. The system enhances pose robustness through a motion-adaptive hybrid tracking module that dynamically switches between Essential matrix and PnP solvers, employs a condition-triggered three-stage optimization pipeline, and leverages foundation models for reliable relocalization. Additionally, it incorporates a Gaussian lifecycle management mechanism based on probabilistic initialization and block-wise multi-view addition/removal, effectively preserving high-frequency scene details while controlling memory usage. Experiments demonstrate that KiloGS-SLAM achieves state-of-the-art accuracy and rendering quality on three large-scale outdoor datasets, efficiently processing sequences exceeding 10,000 frames on a single GPU and successfully reconstructing scenes at the kilometer scale.

Abstract

Scaling monocular 3D Gaussian Splatting (3DGS) SLAM to kilometer-level outdoor environments poses two tightly coupled challenges: fragile long-term pose tracking and excessive memory overhead during large-scale mapping. In this paper, we propose KiloGS-SLAM, a highly efficient and robust monocular 3DGS-SLAM system that jointly addresses both bottlenecks. Since high-fidelity scene reconstruction fundamentally relies on drift-free camera poses, we first introduce a motion-adaptive hybrid tracking module. This module features a condition-triggered three-tier solving pipeline. It dynamically switches between Essential matrix and PnP models to handle geometric degeneracies. An on-demand foundation model can also be activated to rescue the trajectory from catastrophic drift. To ensure the system can sustain these long trajectories without memory exhaustion, we subsequently design a lifecycle-managed Gaussian mapping strategy. By integrating probabilistic initialization with chunk-based multi-view densification and pruning, this full-pipeline optimization effectively reduces primitive redundancy while preserving high-frequency details. Together, the robust tracking guarantees the geometric foundation required for accurate mapping, while the memory-efficient lifecycle-managed mapping enables large-scale operation. Extensive experiments across three challenging outdoor datasets demonstrate that our approach achieves state-of-the-art tracking accuracy and rendering quality, successfully scaling to sequences of over 10,000 frames on a single GPU.

Problem

• monocular SLAM
• 3D Gaussian Splatting
• large-scale mapping
• pose tracking
• memory overhead

Innovation

• 3D Gaussian Splatting
• monocular SLAM
• motion-adaptive tracking
• lifecycle-managed mapping
• large-scale outdoor reconstruction