High-Quality Spatial Reconstruction and Orthoimage Generation Using Efficient 2D Gaussian Splatting

2025-03-25

AI Summary

Traditional TDOM generation relies on DSM construction and explicit occlusion detection, resulting in high computational cost and poor robustness. This paper proposes the first end-to-end TDOM generation paradigm based on 2D Gaussian Splatting (2DGS), completely eliminating DSM modeling and occlusion detection. Our method leverages depth maps to guide 3D scene reconstruction and introduces a divide-and-conquer rasterization scheme for training and rendering, achieving high fidelity in modeling complex terrain and slender structures while significantly improving efficiency. Evaluated on large-scale real-world datasets, our approach achieves superior geometric reconstruction accuracy, produces ortho-imagery of higher quality than conventional methods, accelerates rendering speed substantially, and reduces memory consumption. This work establishes an efficient, robust, and scalable technical pathway for city-scale photogrammetric 3D modeling and downstream decision-support applications.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Data transparency and provenance
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Highly accurate geometric precision and dense image features characterize True Digital Orthophoto Maps (TDOMs), which are in great demand for applications such as urban planning, infrastructure management, and environmental monitoring. Traditional TDOM generation methods need sophisticated processes, such as Digital Surface Models (DSM) and occlusion detection, which are computationally expensive and prone to errors. This work presents an alternative technique rooted in 2D Gaussian Splatting (2DGS), free of explicit DSM and occlusion detection. With depth map generation, spatial information for every pixel within the TDOM is retrieved and can reconstruct the scene with high precision. Divide-and-conquer strategy achieves excellent GS training and rendering with high-resolution TDOMs at a lower resource cost, which preserves higher quality of rendering on complex terrain and thin structure without a decrease in efficiency. Experimental results demonstrate the efficiency of large-scale scene reconstruction and high-precision terrain modeling. This approach provides accurate spatial data, which assists users in better planning and decision-making based on maps.

Problem

• Efficiently generates high-quality orthoimages without DSM
• Improves spatial reconstruction accuracy using 2D Gaussian Splatting
• Reduces computational cost in large-scale terrain modeling

Innovation

• Uses 2D Gaussian Splatting for reconstruction
• Eliminates need for DSM and occlusion detection
• Divide-and-conquer strategy reduces resource cost