arXiv 2310.047872023-12-15cs.ROcs.CV

HI-SLAM：基于混合隐式场的单目实时稠密建图

HI-SLAM: Monocular Real-time Dense Mapping with Hybrid Implicit Fields

University of Stuttgart（斯图加特大学）；Huawei Munich Research Center（华为慕尼黑研究中心）；Central Media Technology Institute, Huawei 2012 Laboratories（华为2012实验室中央媒体技术研究院）；Technical University of Munich（慕尼黑工业大学）

Wei Zhang, Tiecheng Sun, Sen Wang, Qing Cheng, Norbert Haala

备注：: Accepted by IEEE Robotics and Automation Letters

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对现有神经建图框架依赖特殊输入或无法实时运行的问题，提出融合稠密SLAM与神经隐式场的HI-SLAM框架，通过多分辨率编码、Sim(3)位姿图优化及联合深度尺度调整模块，在保证实时性的同时提升建图精度与完整性。

摘要：

在这篇论文中，我们提出了一种基于神经场的实时单目建图框架，用于实现高精度稠密同时定位与建图（Simultaneous Localization and Mapping, SLAM）。现有神经建图框架虽展现出良好效果，但依赖RGB-D输入或位姿输入，或无法实时运行。为解决这些局限，我们的方法将稠密SLAM与神经隐式场相结合。具体而言，我们的稠密SLAM方法并行运行跟踪与全局优化，同时基于最新的SLAM估计结果增量式构建基于神经场的地图。为高效构建神经场，我们采用多分辨率网格编码与有符号距离函数（signed distance function, SDF）表示，这使得地图可始终保持更新，并能通过回环检测即时适配全局更新。为保证全局一致性，我们提出了一种高效的基于Sim(3)的位姿图光束法平差（pose graph bundle adjustment, PGBA）方法，用于实现在线回环检测并缓解位姿与尺度漂移。为进一步提升深度精度，我们引入了学习得到的单目深度先验，并提出了一种新型联合深度与尺度调整（joint depth and scale adjustment, JDSA）模块，以解决深度先验固有的尺度模糊问题。在合成数据集与真实世界数据集上的大量评估验证表明，我们的方法在保持实时性能的同时，在精度与地图完整性方面优于现有方法。

英文摘要：

In this letter, we present a neural field-based real-time monocular mapping framework for accurate and dense Simultaneous Localization and Mapping (SLAM). Recent neural mapping frameworks show promising results, but rely on RGB-D or pose inputs, or cannot run in real-time. To address these limitations, our approach integrates dense-SLAM with neural implicit fields. Specifically, our dense SLAM approach runs parallel tracking and global optimization, while a neural field-based map is constructed incrementally based on the latest SLAM estimates. For the efficient construction of neural fields, we employ multi-resolution grid encoding and signed distance function (SDF) representation. This allows us to keep the map always up-to-date and adapt instantly to global updates via loop closing. For global consistency, we propose an efficient Sim(3)-based pose graph bundle adjustment (PGBA) approach to run online loop closing and mitigate the pose and scale drift. To enhance depth accuracy further, we incorporate learned monocular depth priors. We propose a novel joint depth and scale adjustment (JDSA) module to solve the scale ambiguity inherent in depth priors. Extensive evaluations across synthetic and real-world datasets validate that our approach outperforms existing methods in accuracy and map completeness while preserving real-time performance.

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。