Counting cars, bicycles and pedestrians in traffic in real time, at low cost and in compliance with strict data protection regulations – newer methods of machine learning make this possible. However, empirically valid and meaningful evaluations of such systems in everyday urban life are still lacking.
Municipal traffic censuses are important sources of data for a wide variety of urban actors, especially in the field of urban and traffic planning. Due to existing uncertainties, for example with regard to data protection, municipalities often fall back on established procedures and traffic counts are therefore usually carried out manually with a high use of resources.
The Technologiestiftung Berlin and the Berlin University of Economics and Technology (HTW) have tested the OpenDataCam, which is based on open source, at various locations in the city of Berlin for use in traffic counts.
The Open Data Cam works with machine learning algorithms and makes it possible to carry out accurate traffic counts in real time without saving images. The project was funded by the Federal Ministry of Transport and Digital Infrastructure as part of the mFUND innovation initiative.
In coorporation with: