This is the implementation part of the roofline extraction of the paper: Unsupervised Roofline Extraction from True Orthophotos. First, we use a true orthophoto as input. Next, we perform line extraction to partition the building footprint, which generates separate roof parts. We then utilize a dense point cloud to extrude the partition results and create a LoD2 building model. For generating LoD2 building models, please refer to the repository gfp-building-reconstruction.
If you use it in a scientific work, we kindly ask you to cite it:
PDF BibTeX
@article{sum2021,
author = {Weixiao Gao, Ravi Peters, and Jantien Stoter},
title = {Unsupervised Roofline Extraction from True Orthophotos for LoD2 Building Model Reconstruction},
journal = {Lecture Notes in Geoinformation and Cartography (LNG&C) series},
year={2023},
publisher = {Springer},
}
1. Install all required Python packages as follows
pip install numpy shapely rtree tqdm joblib GDAL
The code was tested on numpy 1.24.2, shapely 2.0.1, rtree 1.0.1, tqdm 4.65.0, joblib 1.2.0, GDAL 3.6.2.
2. Install GCC, Boost (1.63.0 or newer), Eigen3, CGAL, and OpenCV in Conda:
conda install -c conda-forge gcc=12.1.0; conda install -c anaconda boost; conda install -c omnia eigen3; conda install eigen; conda install -c conda-forge cgal; conda install -c conda-forge opencv
The code was tested on Eigen 3.3.7, CGAL 5.5.2, OpenCV 4.7.0.
3. Compile the libkinetic_partition. libraries:
CONDAENV=YOUR_CONDA_ENVIRONMENT_LOCATION
cd kinetic_partition
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release -DPYTHON_LIBRARY=$CONDAENV/lib/libpython3.10.so -DPYTHON_INCLUDE_DIR=$CONDAENV/include/python3.10 -DBOOST_INCLUDE_DIR=$CONDAENV/include -DEIGEN3_INCLUDE_DIR=$CONDAENV/include/eigen3 -DGMP_INCLUDE_DIR=$CONDAENV/include -DGMP_LIBRARY_RELEASE=$CONDAENV/lib/libgmp.so -DMPFR_INCLUDE_DIR=$CONDAENV/include -DMPFR_LIBRARIES=$CONDAENV/lib/libmpfr.so -DGMPXX_INCLUDE_DIR=$CONDAENV/include -DGMPXX_LIBRARIES=$CONDAENV/lib/libgmpxx.so -DCGAL_DIR=$CONDAENV/lib/cmake/CGAL -DOpenCV_DIR=$CONDAENV/lib/cmake/opencv4
make
The code was tested on Ubuntu 22.04 with Python 3.10.
The data in the folder is organized as follows:
data/img/*.tif #orthophotos
data/poly/*.gpkg #building footprints
...
The rest of the folders will be created automatically and the output rooflines will be stored in data/rooflines/*.gpkg .
source activate YOUR_CONDA_ENVIRONMENT
python3 main.py --ROOT_PATH=YOUR_DATA_LOCATION
This implementation is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License or (at your option) any later version. The full text of the license can be found in the accompanying 'License' file.
If you have any questions, comments, or suggestions, please contact me at gaoweixiaocuhk@gmail.com
May. 1st, 2023