📡 Embodied AI Radar具身智能雷达

Topics主题 › Perception, SLAM & 3D Mapping感知、SLAM 与三维建图

Paper论文 2021 · Princeton

DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras

End-to-end learned visual SLAM — recurrent optical-flow updates plus a differentiable bundle-adjustment layer.端到端学习的视觉 SLAM:循环光流更新加上可微分的光束法平差层。

BibTeX
@misc{teed2021droid,
  title = {{DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras}},
  author = {Zachary Teed and Jia Deng},
  year = {2021},
  eprint = {2108.10869},
  archivePrefix = {arXiv},
  primaryClass = {cs.CV},
  url = {https://arxiv.org/abs/2108.10869},
}

Authors (from arXiv): 作者(来自 arXiv):Zachary Teed, Jia Deng

More from: 同机构更多条目:Princeton普林斯顿大学

More in Perception, SLAM & 3D Mapping更多感知、SLAM 与三维建图条目

All Perception, SLAM & 3D Mapping entries →查看全部感知、SLAM 与三维建图条目 →