NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields. https://arxiv.org/abs/2210.13641 + Sigma-Fusion: Probabilistic Volumetric Fusion for Dense Monocular SLAM https://arxiv.org/abs/2210.01276 - NeRF-SLAM/gui/open3d_gui.py at master
NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields. https://arxiv.org/abs/2210.13641 + Sigma-Fusion: Probabilistic Volumetric Fusion for Dense Monocular SLAM https://arxiv.org/abs/2210.01276 - NeRF-SLAM/fusion/fusion_module.py at
git clone https://github.com/ToniRV/NeRF-SLAM.git --recurse-submodules git submodule update --init --recursive From this point on, use a virtual environment... Install torch (seeherefor other versions): # CUDA 11.3 pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 --extra-ind...
代码链接:https://github.com/ToniRV/NeRF-SLAM NeRF-SLAM实际是DROID-SLAM+probabilistic volumetric fusion+Instant NGP三个方案的组合,主要创新在于引入了深度和位姿的不确定性。 这篇文章希望解决什么问题? 单目稠密SLAM也好,直接进行单目深度估计也好,得到的深度图很多数值是不能用的,也很自然的不能用于三角化等...
代码链接:https://github.com/chenhsuanlin/bundle-adjusting-NeRF 官方主页:https://chenhsuanlin.bitbucket.io/bundle-adjusting-NeRF/ 这篇文章希望解决什么问题? 传统的NeRF训练需要非常精确的位姿,并且NeRF SLAM反向传播优化位姿时,对位姿初值很敏感,容易陷入局部最优。
NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields. https://arxiv.org/abs/2210.13641 + Sigma-Fusion: Probabilistic Volumetric Fusion for Dense Monocular SLAM https://arxiv.org/abs/2210.01276 - NeRF-SLAM/utils/open3d_pickle.py at m
代码链接:https://github.com/kxhit/vMAP 官方主页:https://kxhit.github.io/vMAP 也是一篇开源的物体级NeRF SLAM方案,同样是为每个目标单独分配一个MLP。 这篇文章希望解决什么问题? 物体级地图不够稠密。 核心思想是什么? 每个目标都单独训练一个MLP,建立物体级NeRF地图。
NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields. https://arxiv.org/abs/2210.13641 + Sigma-Fusion: Probabilistic Volumetric Fusion for Dense Monocular SLAM https://arxiv.org/abs/2210.01276 - NeRF-SLAM/networks/droid_frontend.py
Real-Time Dense Monocular SLAM with Neural Radiance Fields. https://arxiv.org/abs/2210.13641 + Sigma-Fusion: Probabilistic Volumetric Fusion for Dense Monocular SLAM https://arxiv.org/abs/2210.01276 - NeRF-SLAM/requirements.txt at master · tccoin/N
代码链接:https://github.com/chenhsuanlin/bundle-adjusting-NeRF 官方主页:https://chenhsuanlin.bitbucket.io/bundle-adjusting-NeRF/ 这篇文章希望解决什么问题? 传统的NeRF训练需要非常精确的位姿,并且NeRF SLAM反向传播优化位姿时,对位姿初值很敏感,容易陷入局部最优。