Count cars, bicycles and pedestrians in real time on the road, at low cost and in compliance with data protection - newer machine learning methods make this possible. So far, however, there is still a lack of empirically valid and meaningful evaluations of corresponding systems in everyday urban life. OpenTrafficCount tests the application of a counting system developed from open source components at various locations in the Berlin city area. In coordination with the Berlin traffic administration, the results will be compared with random samples of manual traffic counts.
The system to be tested is characterised by low costs, variable application possibilities, real-time evaluation and data protection conformity. This results in applications for urban planning, companies, but also for civil society (e.g. Citizen Science). With this system, temporary changes in the traffic area can be analysed relatively easily, the effects of road works can be investigated or danger zones can be observed in more detail. The collected data is compiled in a cloud-based database in compliance with data protection regulations and made available under a free licence via standardised interfaces for administration, business and civil society.