Forschung

SIDING – Rail Connection Identification through Intelligent Geolocation

An important measure to reduce CO₂ emissions in Germany is the shift of freight transport to rail. Manufacturing companies require access points to the Deutsche Bahn rail network for this purpose. Many such rail sidings are located on private company premises and are not listed in any central registry. The SIDING project uses satellite imagery and automated image recognition via machine learning to detect private rail sidings from aerial views. In this way, a comprehensive dataset for Germany is being created, enabling companies to identify access points to the rail network and shift their freight to rail.

The methods and results of this project were presented at the 2023 Dresden Land Use Symposium. The presentation is available here: https://zenodo.org/record/8095902