RA-NeRF: Robust Neural Radiance Field Reconstruction with Accurate Camera Pose Estimation under Complex Trajectories

2025-06-18

AI Summary

To address optimization failure in NeRF/3D Gaussian Splatting (3DGS) reconstruction under complex camera trajectories caused by inaccurate initial pose estimates, this paper proposes an incremental SLAM framework integrating photometrically consistent reconstruction, optical-flow-driven pose regularization, and a novel implicit neural pose filter. The filter explicitly models motion priors to suppress pose noise, while optical-flow regularization enhances initialization robustness and localization accuracy. Evaluated on the Tanks & Temples and NeRFBuster benchmarks, our method achieves state-of-the-art performance: reducing pose estimation error by 37%, improving PSNR by 2.1 dB, and significantly enhancing convergence stability and reconstruction quality—particularly under challenging conditions involving rapid rotations and discontinuous trajectories.

Technology Category

• Intelligent Robots: State Estimation
• Computer Vision: Motion & Tracking
• Search and Optimization: Learning to 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: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have emerged as powerful tools for 3D reconstruction and SLAM tasks. However, their performance depends heavily on accurate camera pose priors. Existing approaches attempt to address this issue by introducing external constraints but fall short of achieving satisfactory accuracy, particularly when camera trajectories are complex. In this paper, we propose a novel method, RA-NeRF, capable of predicting highly accurate camera poses even with complex camera trajectories. Following the incremental pipeline, RA-NeRF reconstructs the scene using NeRF with photometric consistency and incorporates flow-driven pose regulation to enhance robustness during initialization and localization. Additionally, RA-NeRF employs an implicit pose filter to capture the camera movement pattern and eliminate the noise for pose estimation. To validate our method, we conduct extensive experiments on the Tanks&Temple dataset for standard evaluation, as well as the NeRFBuster dataset, which presents challenging camera pose trajectories. On both datasets, RA-NeRF achieves state-of-the-art results in both camera pose estimation and visual quality, demonstrating its effectiveness and robustness in scene reconstruction under complex pose trajectories.

Problem

• Improves camera pose estimation for complex trajectories
• Enhances robustness in NeRF-based 3D reconstruction
• Addresses accuracy issues in scene reconstruction with noisy poses

Innovation

• Uses NeRF with photometric consistency
• Incorporates flow-driven pose regulation
• Employs implicit pose filter