Enabled by the ubiquitous generation of geo-referenced tracking data, there has been a recent surge of algorithms that construct street maps from tracking data.
However, visual inspection is still the most common approach to evaluate the quality of the algorithms, and cross-comparison of different algorithms is very rare, since algorithms and constructed maps are generally not publicly available. There is a lack of benchmark data, and quantitative evaluation with suitable distance measures has just recently begun.
This repository will provide access to
- a set of map generation algorithms,
- data sets including tracking data as well as respective map datasets, and
- various distance measures to assess the quality of the constructed maps.
As they become available, we encourage interested scientists and developers to add their source code to this site.
Should you use the source code and/or data from this site, please cite also the following paper:
M. Ahmed, S. Karagiorgou, D. Pfoser, and C. Wenk. A Comparison and Evaluation of Map Construction Algorithms. GeoInformatica, 19(3):601-632, 2015.
DOI: 10.1007/s10707-014-0222-6 - paper
Refer to the project Web page for more details.