融合学习与优化，实现相机晃动下的高精度实时稠密重建
PROFusion: Robust and Accurate Dense Reconstruction via Camera Pose Regression and Optimization
- 用网络预测相对位姿作为优化起点，提升初始化鲁棒性
- 在复杂运动下优于最强基线，在稳定运动中保持同等精度
- 适合需要高鲁棒性与精度的机器人场景重建任务
实时稠密场景重建对机器人应用至关重要，但现有RGB-D SLAM系统在相机大幅视角变化、快速运动或剧烈抖动时表现不佳。传统基于优化的方法精度高但对初始化敏感，而基于学习的方法虽鲁棒却难以满足稠密重建的精度要求。本文提出一种结合学习初始化与优化精化的方案：通过相机位姿回归网络从连续的RGB-D帧中预测度量感知的相对位姿，作为随机优化算法的可靠初始值，进一步对齐深度图与场景几何。大量实验表明，该方法在挑战性基准上超越最优对比模型，同时在稳定运动序列上保持相当精度。系统支持实时运行，证明简单且原理清晰的技术组合可兼顾不稳定运动下的鲁棒性与稠密重建的准确性。代码已开源：https://github.com/siyandong/PROFusion。
原文摘要 · Abstract (English)
Real-time dense scene reconstruction during unstable camera motions is crucial for robotics, yet current RGB-D SLAM systems fail when cameras experience large viewpoint changes, fast motions, or sudden shaking. Classical optimization-based methods deliver high accuracy but fail with poor initialization during large motions, while learning-based approaches provide robustness but lack sufficient accuracy for dense reconstruction. We address this challenge through a combination of learning-based initialization with optimization-based refinement. Our method employs a camera pose regression network to predict metric-aware relative poses from consecutive RGB-D frames, which serve as reliable starting points for a randomized optimization algorithm that further aligns depth images with the scene geometry. Extensive experiments demonstrate promising results: our approach outperforms the best competitor on challenging benchmarks, while maintaining comparable accuracy on stable motion sequences. The system operates in real-time, showcasing that combining simple and principled techniques can achieve both robustness for unstable motions and accuracy for dense reconstruction. Code released: https://github.com/siyandong/PROFusion.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。