ULSR-GS: Ultra Large-scale Surface Reconstruction Gaussian Splatting with Multi-View Geometric Consistency

2024-12-02

arXiv.org

AI Summary

To address the poor scale adaptability and low detail fidelity of Gaussian Splatting (GS) in large-scale aerial image-based surface reconstruction, this paper proposes ULSR-GS—a unified large-scale reconstruction framework. Methodologically, it introduces: (1) a novel point-to-image partitioning strategy coupled with multi-view optimal view matching to alleviate global optimization bottlenecks; (2) a dynamic densification scheme grounded in multi-view geometric consistency, enabling adaptive enhancement of point cloud density and structural coherence over complex urban terrain; and (3) a hybrid training paradigm integrating Gaussian rendering with region-wise collaborative optimization. Evaluated on large-scale aerial photogrammetry benchmarks, ULSR-GS substantially outperforms existing GS-based methods, achieving marked improvements in surface extraction accuracy. It is the first GS-based approach to enable high-resolution, high-fidelity 3D surface reconstruction at the square-kilometer scale.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

While Gaussian Splatting (GS) demonstrates efficient and high-quality scene rendering and small area surface extraction ability, it falls short in handling large-scale aerial image surface extraction tasks. To overcome this, we present ULSR-GS, a framework dedicated to high-fidelity surface extraction in ultra-large-scale scenes, addressing the limitations of existing GS-based mesh extraction methods. Specifically, we propose a point-to-photo partitioning approach combined with a multi-view optimal view matching principle to select the best training images for each sub-region. Additionally, during training, ULSR-GS employs a densification strategy based on multi-view geometric consistency to enhance surface extraction details. Experimental results demonstrate that ULSR-GS outperforms other state-of-the-art GS-based works on large-scale aerial photogrammetry benchmark datasets, significantly improving surface extraction accuracy in complex urban environments. Project page: https://ulsrgs.github.io.

Problem

• Enables large-scale aerial surface extraction with Gaussian Splatting
• Improves surface detail via multi-view geometric consistency
• Optimizes training image selection for sub-region fidelity

Innovation

• Point-to-photo partitioning for sub-region training
• Multi-view geometric consistency densification strategy
• Optimal view matching for image selection