Dense3DReconstruction

Project Brief

This is official repository for Project for Real-Time Globally Consistent Dense 3D Reconstruction With Online Texturing

This is a project from LuVision SIGMA, Tsinghua University. Visit our website for more interesting works: http://www.luvision.net/

Related papers:

1. Multi-Index Hashing for Loop closure Detection. International Conference on Multimedia Expo, 2017. Best Student Paper Awards.

2. Beyond SIFT Using Binary features in Loop Closure Detection. IROS 2017.

3. FlashFusion: Real-time Globally Consistent Dense 3D Reconstruction using CPU Computing. RSS 2018.

4. Real-Time Globally Consistent Dense 3D Reconstruction With Online Texturing. TPAMI 2020.

Please include citation below if you find our project helpful.

@article{Dense3DReconstruction,
  author={Han, Lei and Gu, Siyuan and Zhong, Dawei and Quan, Shuxue and Fang, Lu},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence}, 
  title={Real-Time Globally Consistent Dense 3D Reconstruction With Online Texturing}, 
  year={2022},
  volume={44},
  number={3},
  pages={1519-1533},
  doi={10.1109/TPAMI.2020.3021023}}

License

This project is released under the GPLv3 license. We only allow free use for academic use. For commercial use, please contact us to negotiate a different license by: fanglu at tsinghua.edu.cn

Setup Instruction

0. Set up mirror address to accelerate installation

1. Install basic build tools:

sudo apt-get install build-essential pkg-config cmake git

2. Install required library step by step:

• First-to-install

sudo apt-get install ffmpeg libsuitesparse-dev libxmu-dev libgtk-3-dev libv4l-dev libglew-dev qtbase5-dev

• Glut

sudo apt-get install freeglut3-dev

• boost

sudo apt-get install libboost-all-dev

• Eigen3

sudo apt-get install libeigen3-dev

• Sophus

git clone https://github.com/strasdat/Sophus.git
cd Sophus && mkdir build && cd build
cmake .. && make -j3 && sudo make install

• Pangolin

git clone https://github.com/stevenlovegrove/Pangolin.git
cd Pangolin && mkdir build && cd build
cmake .. && make -j3 && sudo make install

• OpenNI2

sudo apt-get install libopenni2-dev

• librealsense (to use realsense, check Intel's website to install essential driver)

sudo apt-get install libfreenect-dev libglfw3-dev
git clone https://github.com/IntelRealSense/librealsense.git
cd librealsense && mkdir build && cd build
cmake .. && make -j3 && sudo make install

• OpenCV

sudo apt-get install libopencv-dev libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev
git clone https://github.com/opencv/opencv.git
cd opencv && git checkout 8f1356c && mkdir build && cd build
cmake -DCMAKE_BUILD_TYPE=RELEASE -DCMAKE_INSTALL_PREFIX=/usr/local -DWITH_TBB=ON -DWITH_LIBV4L=ON
-DWITH_QT=ON -DWITH_OPENGL=ON -DWITH_CUDA=OFF -DBUILD_PERF_TESTS=OFF -DBUILD_TESTS=OFF ..
make -j3 && sudo make install
sudo /bin/bash -c 'echo "/usr/local/lib" > /etc/ld.so.conf.d/opencv.conf' 
sudo ldconfig

3. Compile this repository

git clone https://github.com/THU-luvision/Dense3DReconstruction.git
cd Dense3DReconstruction && mkdir build && cd build
cmake .. &&make -j3

Usage Instruction

./FlashFusion $DataFolder $ParametersFile Resolution inputID

For each input parameter:

• $DataFolder: base folder containing test data or output data.

The basefolder should contain:

1. calib.txt for calibration parameters
2. associate.txt for input rgbd images
3. groundtruth.txt for TUM RGBD dataset groundtruth data.

An example datafolder is provided here and here

• $ParametersFile: input parameters, an example is given as settings.yaml

• Resolution: Reconstruction voxel resolution, ranges from 0.005 to 0.04

• InputMode:

  0: offline data stored in DataFolder 1: Online data using ASUS Xtion 2: Online data using Intel Realsense

Example

cd $SourceDirectory/build
./FlashFusion  ~/Downloads/xtion/ ../settings.yaml 0.005 0