A-TDOM: Active TDOM via On-the-Fly 3DGS
🤖 AI Summary
Traditional TDOM generation relies on offline photogrammetric pipelines, suffering from high latency, inaccurate camera pose estimation, DSM inaccuracies, and occlusion-induced quality instability. This paper proposes a near-real-time TDOM generation framework that uniquely integrates on-the-fly Structure-from-Motion (SfM) with dynamic 3D Gaussian Splatting (3DGS) optimization, enabling incremental Gaussian fusion and orthographic splatting rendering. Upon arrival of a new image, the system completes pose estimation, sparse point cloud reconstruction, dynamic Gaussian field refinement, and orthorectified rendering within seconds—without requiring global re-optimization. Evaluated on multiple benchmark datasets, the method achieves a favorable trade-off between geometric accuracy and visual fidelity, significantly improving both timeliness and robustness of TDOM generation. It establishes a novel paradigm for real-time applications such as UAV-based inspection and emergency surveying.
Application Category
📝 Abstract
True Digital Orthophoto Map (TDOM) serves as a crucial geospatial product in various fields such as urban management, city planning, land surveying, etc. However, traditional TDOM generation methods generally rely on a complex offline photogrammetric pipeline, resulting in delays that hinder real-time applications. Moreover, the quality of TDOM may degrade due to various challenges, such as inaccurate camera poses or Digital Surface Model (DSM) and scene occlusions. To address these challenges, this work introduces A-TDOM, a near real-time TDOM generation method based on On-the-Fly 3DGS optimization. As each image is acquired, its pose and sparse point cloud are computed via On-the-Fly SfM. Then new Gaussians are integrated and optimized into previously unseen or coarsely reconstructed regions. By integrating with orthogonal splatting, A-TDOM can render just after each update of a new 3DGS field. Initial experiments on multiple benchmarks show that the proposed A-TDOM is capable of actively rendering TDOM in near real-time, with 3DGS optimization for each new image in seconds while maintaining acceptable rendering quality and TDOM geometric accuracy.
Problem
Research questions and friction points this paper is trying to address.
Real-time True Digital Orthophoto Map generation challenges
Overcoming inaccurate camera poses and scene occlusions
Reducing delays from traditional offline photogrammetric pipelines
Innovation
Methods, ideas, or system contributions that make the work stand out.
On-the-Fly 3DGS optimization method
Real-time Gaussian integration and optimization
Orthogonal splatting for instant TDOM rendering
🔎 Similar Papers
No similar papers found.
