From 2024, Belgian public transport company De Lijn counts the number of passengers boarding and disembarking on its vehicles using a smart AI algorithm that combines ticket registrations with counts by 'people counting cameras'. At the beginning of May, De Lijn had already installed 350 cameras on its vehicles. By the end of this year, 1,600 buses and trams will have this new equipment on board.
![De LIJN AI](/sites/europe/files/2024-05/Site%2023-5-24%20DeLijn%20AI%2020240521_142956.jpg)
A smart AI algorithm links the counting data from the cameras with the existing registrations via the issued tickets. The algorithm takes into account a number of factors such as type of vehicle, weather or type of day (school day or weekend day for example) or type of trip (peak or off-peak) and even the number of online stop requests. The new counting method therefore takes many more factors into account. According to De Lijn, the new counting method also shows that anyone who does not scan their ticket will from now on end up in the transport statistics. It offers De Lijn richer and better data about its passenger flows. This allows the network to be even more responsive to demand. De Lijn started a test on 15 vehicles with such 3D or people counting cameras in November 2022. The cameras proved to be effective and counted the number of passengers arriving and departing with up to 95% accuracy. Until the end of last year, De Lijn determined its passenger numbers based on two sources. Firstly, the scans with subscriptions, physical tickets and payments by bank card on the vehicles were used. Secondly, the tickets that are not scanned (season tickets and SMS tickets) were allocated to the different journeys and lines in proportion to the distribution of the effective scans of physical tickets. Regular research was also conducted into the scanning behaviour of subscribers. Finally, the algorithm used to determine passenger numbers took into account a so-called 'increment factor'.