This code implements the Incremental Reconstruction algorithm described in the paper Piazza, Enrico, Andrea Romanoni, and Matteo Matteucci. "Real-Time CPU-Based Large-Scale Three-Dimensional Mesh Reconstruction." IEEE Robotics and Automation Letters 3.3 (2018): 1584-1591.

Bibtex:

@article{piazza2018real,
  title={Real-Time CPU-Based Large-Scale Three-Dimensional Mesh Reconstruction},
  author={Piazza, Enrico and Romanoni, Andrea and Matteucci, Matteo},
  journal={IEEE Robotics and Automation Letters},
  volume={3},
  number={3},
  pages={1584--1591},
  year={2018},
  publisher={IEEE}
}

DEPENDENCIES

• opencv
• eigen3
• gmp
• mpfr
• CGAL
• boost

If you are on Ubuntu, you can install all the dependencies (execept CGAL) with:

sudo apt install libopencv-dev libeigen3-dev libgmp-dev \
libmpfr-dev libmpfr4 libboost-all-dev

since CGAL is not up to date on ubuntu distro we advise to install it from the sources.

Generate runnable files

You can compile the sources using the simple cmake-make procedure. First thing first, create the build directory:

mkdir build

Then enter the directory and generate the make file and the executables:

cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
make -j4

Let notice, in the external directory the sources of glm and rapidjson are provided, therefore they do not need to be installed. However, if you already have those installed on your machine you may want to use your version. To do so:

1. open the CMakeLists.txt file
2. comment out lines 20 and 21
3. add find_library(GMP_LIBRARY gmp /usr/lib) between line 9 and 10.

Run

Once the executable files have been generated you would find sfmReconstructor and slamReconstructor files into the build folder. To run one of them, e.g., sfmReconstructor, do the following:

1. go to the project main directory
2. create the output folder (you need to do this only once)

mkdir output

3. launch your file specifing the input data

sfmReconstructor data/sfm_data.json