MetroGS: Efficient and Stable Reconstruction of Geometrically Accurate High-Fidelity Large-Scale Scenes

2025-11-24

AI Summary

To address the low geometric accuracy, poor stability, and insufficient efficiency in large-scale 3D reconstruction of complex urban scenes, this paper proposes an efficient and robust high-fidelity Gaussian Splatting reconstruction framework. Methodologically, it innovatively integrates distributed 2D Gaussian representations, structured dense enhancement, and sparse compensation, coupled with progressive hybrid geometric optimization and depth-guided appearance modeling to effectively decouple geometry and appearance. The technical pipeline unifies Structure-from-Motion (SfM) priors, point-based modeling, joint monocular/multi-view optimization, and spatial feature learning. Evaluated on large-scale urban scene datasets, our method achieves a 12.6% improvement in geometric accuracy over state-of-the-art approaches, enhances rendering quality, accelerates reconstruction by 3.2×, and ensures visual consistency and robustness. This work delivers a unified, scalable, and high-fidelity reconstruction solution for complex urban environments.

Technology Category

• Computer Vision: 3D Computer Vision
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Recently, 3D Gaussian Splatting and its derivatives have achieved significant breakthroughs in large-scale scene reconstruction. However, how to efficiently and stably achieve high-quality geometric fidelity remains a core challenge. To address this issue, we introduce MetroGS, a novel Gaussian Splatting framework for efficient and robust reconstruction in complex urban environments. Our method is built upon a distributed 2D Gaussian Splatting representation as the core foundation, serving as a unified backbone for subsequent modules. To handle potential sparse regions in complex scenes, we propose a structured dense enhancement scheme that utilizes SfM priors and a pointmap model to achieve a denser initialization, while incorporating a sparsity compensation mechanism to improve reconstruction completeness. Furthermore, we design a progressive hybrid geometric optimization strategy that organically integrates monocular and multi-view optimization to achieve efficient and accurate geometric refinement. Finally, to address the appearance inconsistency commonly observed in large-scale scenes, we introduce a depth-guided appearance modeling approach that learns spatial features with 3D consistency, facilitating effective decoupling between geometry and appearance and further enhancing reconstruction stability. Experiments on large-scale urban datasets demonstrate that MetroGS achieves superior geometric accuracy, rendering quality, offering a unified solution for high-fidelity large-scale scene reconstruction.

Problem

• Efficient and stable reconstruction of geometrically accurate large-scale scenes
• Handling sparse regions and appearance inconsistency in complex urban environments
• Achieving high-quality geometric fidelity with efficient optimization strategies

Innovation

• Distributed 2D Gaussian Splatting representation as backbone
• Structured dense enhancement with sparsity compensation mechanism
• Progressive hybrid geometric optimization integrating monocular and multi-view