arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2205.01656cs.CV
收藏

GeoRefine:用于精确稠密映射的自监督在线深度精化

GeoRefine: Self-Supervised Online Depth Refinement for Accurate Dense Mapping

OPPO US Research Center, InnoPeak Technology(OPPO美国研究中心,InnoPeak科技)

Pan Ji, Qingan Yan, Yuxin Ma, Yi Xu

摘要:

我们提出了一个强大的和准确的深度细化系统,名为GeoRefine,几何一致的密集映射从单目序列。GeoRefine由三个模块组成:使用基于学习的先验知识的混合SLAM模块,利用自我监督的在线深度细化模块,以及通过TSDF融合的全局映射模块。所提出的系统通过设计是在线的,并且通过以下方式实现了很大的鲁棒性和准确性:(i)结合了基于学习的光流和/或深度的鲁棒混合SLAM;(ii)利用SLAM输出并强制长期几何一致性的自监督损失;(iii)避免在线深度细化中的退化情况的仔细的系统设计。我们在多个公共数据集上对GeoRefine进行了广泛的评估,并达到了$5\%$绝对相对深度误差的低水平。

英文摘要:

We present a robust and accurate depth refinement system, named GeoRefine, for geometrically-consistent dense mapping from monocular sequences. GeoRefine consists of three modules: a hybrid SLAM module using learning-based priors, an online depth refinement module leveraging self-supervision, and a global mapping module via TSDF fusion. The proposed system is online by design and achieves great robustness and accuracy via: (i) a robustified hybrid SLAM that incorporates learning-based optical flow and/or depth; (ii) self-supervised losses that leverage SLAM outputs and enforce long-term geometric consistency; (iii) careful system design that avoids degenerate cases in online depth refinement. We extensively evaluate GeoRefine on multiple public datasets and reach as low as $5\%$ absolute relative depth errors.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑