anaumannigg / PolyMatching

Many-to-Many Polygon Matching à la Jaccard

The code in this repository implements the algorithm proposed in the paper 'Many-to-Many Polygon Matching à la Jaccard'.

Dependencies

The code makes use of the following libraries:

• boost
• CGAL
• shapelib
• gurobi

Installation

1. Clone https://github.com/anaumannigg/m-n-Polygon-Matching-Jaccard.git
2. Clone https://github.com/OSGeo/shapelib.git into the subdirectory 'lib'.
3. Put the FindGUROBI.cmake file into the project folder
4. make project:

mkdir build
cd build
cmake ..
make -j 8

5. Run via

 ./build/PolygonMatching

Running the Code

The code can be parameterized with the following command line arguments:

• -d dataset_name | Provide the name of the dataset.
• -l lambda | Specify the parameter for lambda.
• -t threads | (optional, default 1) Specify the number of threads that will be used for computations on connected components. (We recommend 8)
• -k time_limit | (optional, default is 1000) Specify the time limit in [s] for the computation per connected component.
• -s size_limit | (optional, deactivated on default) Set a limit on the set size on cumulative vertex computation.
• -h, --help | Display help message.

The program expects two Shapefiles containing single Polygons without holes, which are located in 'input/data_name'. The files should be named in the way 'data_name_osm' and 'data_name_atkis'. Should a polygon have holes, only its outer boundary will be considered. The code will store its results in 'export/data_name/lambdavalue*10' in the form of Shapefiles containing the original polygon sets per input file, where each polygon has the attribute 'match_id', which encodes the match, that the polygon has been assigned to. Negative match IDs are used for polygons, which have not been matched. Additionally, two csv-Files are exported into the same directory.

• 'analysis' - provides the sizes of the components after the precomputation steps as well as information about timing, and if the optimal solution could be computed within the given time limit.
• 'data' provides the information about the optimal solution, where for each polygon, its match ID as well as the quality of the match is listed.

Within this table 'map' = 0 refers to the osm data and 'map' = 1 to atkis.

This repository contains one exemplary dataset of Auerberg, a district of Bonn. It consists of the OpenStreeMap-Data, fetched from the Overpass-API and the official cadastral data, publicly available as a free download. (both data sets were fetched on November 9th, 2023)