GitHub is where people build software. More than 100 million people use GitHub to discover, fork, and contribute to over 420 million projects.
1、LIV-GaussMap(https://github.com/sheng00125/LIV-GaussMap),港科大沈邵劼组的成果,用LI里程计初始化GS地图,再用光度梯度优化高斯和球谐,开放了github但是代码暂未开源。 2、MM3DGS SLAM(开源了,https://vita-group.github.io/MM3DGS-SLAM/)、Gaussian-LIC(没开源,https://arxiv.org/pdf/2404.06926),...
# Install our modified code (cuda) git clone git@github.com:npu-yanchi/diff-gaussian-rasterization-for-gsslam.git cd diff-gaussian-rasterization-for-gsslam python setup.py install pip install . Calling method: rendered_image, radii, rendered_depth, rendered_alpha, render_depth_var = rasterizer...
MM3DGS-SLAM The website of our paper for "MM3DGS SLAM: Multi-modal 3D Gaussian Splatting for SLAM Using Vision, Depth, and Inertial Measurements". [Project Page]|[Video]|[Paper]|[Arxiv] Framework As shown above, we present the framework for Multi-modal 3D Gaussian Splatting for SLAM....
# Install our modified code (cuda) git clone git@github.com:npu-yanchi/diff-gaussian-rasterization-for-gsslam.git cd diff-gaussian-rasterization-for-gsslam python setup.py install pip install . Calling method: rendered_image, radii, rendered_depth, rendered_alpha, render_depth_var = rasterizer...
代码链接:gapszju.github.io/RTG-S 2. 摘要 我们提出了 RTG-SLAM,这是一个使用高斯点阵的 RGBD 相机进行大规模环境实时三维重建系统。RTG-SLAM 具有紧凑的高斯表示和高效的即时高斯优化方案。我们强制使每个高斯要么不透明要么几乎透明,其中不透明的高斯适应表面和主导颜色,透明的高斯适应残余颜色。通过以与颜色渲染...
git clone git@github.com:VITA-Group/MM3DGS-SLAM.git --recursive Environment setup The simplest way to install dependencies is withanaconda conda create -n mm3dgs python=3.10 conda activate mm3dgs conda install -c"nvidia/label/cuda-11.8.0"cuda-toolkit conda env update -f environment.yml ...
RGBD-3DGS-SLAM is a monocular SLAM system leveraging 3D Gaussian Splatting (3DGS) for accurate point cloud and visual odometry estimation. By integrating neural networks, it estimates depth and camera intrinsics from RGB images alone, with optional suppo
代码链接:https://github.com/lab-of-ai-and-robotics/gs_icp_slam 2. 摘要 同时定位与地图构建(SLAM)在机器人学、虚拟现实(VR)和增强现实(AR)应用中起着关键作用。近期对于密集表示SLAM的进展突显了利用神经场景表示和3D高斯表示来实现高保真空间表示的潜力。本文提出了一种新颖的密集表示SLAM方法,采用广义迭代最...
代码链接:https://gapszju.github.io/RTG-SLAM 我们提出了 RTG-SLAM,这是一个使用高斯点阵的 RGBD 相机进行大规模环境实时三维重建系统。RTG-SLAM 具有紧凑的高斯表示和高效的即时高斯优化方案。我们强制使每个高斯要么不透明要么几乎透明,其中不透明的高斯适应表面和主导颜色,透明的高斯适应残余颜色。通过以与颜色渲染...