If you cannot measure it, you cannot improve it. With this adagium in mind, Bureau Marineterrein Amsterdam, CITIXL, TAPP, and Amsterdam CTO collaborated to deploy smart cameras and visitor screens, to better manage the growing crowds at Marineterrein Amsterdam.
If you cannot measure it, you cannot improve it. With this adagium in mind, CITIXL, TAPP, and Amsterdam CTO collaborated to deploy smart cameras and visitor screens, to better manage the growing crowds at Marineterrein Amsterdam. The smart cameras count visitors which are anonymized and visualized on on-site screens, providing visitors with live information on the crowdedness per area and the current water quality.
When is a crowded public space too crowded? Is there a turning point at which visitors will no longer feel safe and are unable to move around freely? How do visitors anticipate their next visit when informed about the crowdedness in a public space?
The open-source crowd monitor will be tested for two months, and aims to identify the crowdedness and liveability of Marineterrein Amsterdam during the summer.
Six video cameras and a computer algorithm are used to determine how busy it is on the site. This self-learning algorithm automatically counts visitor numbers and visitor density at six locations: the entrances, the Voorwerf, the inner harbour, and the fitness garden.
Visitors are anonymized before the images are analysed by the algorithm. This feature was developed by the Municipality of Amsterdam and the source code will become freely available soon. Video images are locally processed into data via a secure end-to-end encrypted connection and then deleted.
Personal data is processed carefully and securely according to the applicable legislation (GDPR) and transparency guidelines (TADA). This project is subject to a specific privacy policy and the camera locations and functions are included in the camera register of the Municipality of Amsterdam.
Visitors are informed of the Marineterrein’s busiest and quietest locations via the website, the mobile app, and on-site information screens. The number of people is counted as pixelated bodies entering or occupying the public space, and these defined areas are recognized as ‘events’ by the smart cameras. These ‘events’ are then visualized in the form of interactive data that indicate how many people are using a space at a given time. The data on these cameras is completely anonymous and will not be shared.
This experiment will investigate whether visitors adapt their behaviours to avoid crowds, maintain a distance of 1.5 metres, and ask questions about their privacy. The open source crowd monitor uses new technologies to provide visitors with information while simultaneously safeguarding their privacy. Personal data is processed carefully and securely according to the applicable legislation (GDPR) and transparency guidelines (TADA).
