Economic Services Departments
Automated parking control
TagsParking control Image recognition License plate reading
In Amsterdam, the number of cars allowed to park in the city is limited, keeping the city liveable and accessible. The municipality checks whether a parked car has the right to be parked, for example, because parking fees have been paid via a parking meter or app, or because the owner has a parking permit. Enforcement is done with the help of scan cars equpiped with camera’s, automating the process of license plate identification and background checks with specific scanning equipment and AI-based identification service. The service is currently in use for more than 150,000 street parking spaces in the City of Amsterdam.
The service follows a three-step process. In the first step, the scan cars drive through the city and use object recognition software to scan and identify the license plates of surrounding cars. After the identification, the license plate number is checked against the National Parking Register to validate if the car has permission to park at a given location. Whenever no payment has been made for current parking, the case is sent to a human inspector for further processing. In the last step, the parking inspector assesses the scanned images to determine whether the license plate was recognized correctly and whether there is a special situation such as loading or unloading. The parking inspector may also verify the situation on-site. Whenever there is no valid reason found for non-paid parking, a parking ticket is issued.
- Parking services
- Egis Parking Services B.V.
- +31 20 624 111
More detailed information on the system
Here you can get acquainted with the information used by the system, the operating logic, and its governance in the areas that interest you.
- Datasets Show More Show Less
Key data sources utilised in the development and use of the system, their content and utilisation methods. The different data sources are separated by subheadings.
NameLicense plate scans
Cars equipped with automated license plate recognition drive around the city, scanning the license plates of parked cars. The car’s scanning software finds and isolates the license plates from the camera data stream of the street surroundings. This data collected by the scan cars consists of the scanned images of license plates along with car location and timestamp data. The data is retained for 48 hours for the cases with paid parking fees, and 13 weeks for the cases with unpaid fees.
NameNational Parking Register
The National Parking Register is a national database containing all transactions concerning parking rights in the Netherlands. The register is owned and run by RDW. Data is collected from parking meters and parking applications from municipal and commercial parking companies, parking providers, and law enforcers. The parking rights data is retained (pseudonymized) in the register for 13 weeks. The register’s data is used to check the validity of rights for given parking.
- Human oversight Show More Show Less
Human oversight during the use of the service.
Human parking inspectors play a key role in assessing the scan images to identify whether there is a valid reason for non-paid parking fees. These situations could be related to, for example, loading or unloading, or stationary cars in front of traffic lights. The parking inspectors also assess whether the characters on the license plate were identified correctly. Based on this remote assessment, the inspector can decide whether to conduct an on-site visit to assure the full comprehension of the situation or not.
- Data processing Show More Show Less
The operational logic of the automatic data processing and reasoning performed by the system and the models used.
System architecture description
The service uses license plate recognition algorithms to locate and process the license plate data from the camera data stream. Algorithms are used to locate the license plate from the image data, to adjust the images for identification, to identify the individual characters of the license plate, and to validate the plate contents against national license plate characteristics.
After a successful plate identification and processing, license plate data is sent to the National Parking Register for further processing. NPR’s algorithm checks the validity of parking rights for the license plate in a given time and location (for technical information on the NPR algorithm, see the information on their website: https://nationaalparkeerregister.nl/fileadmin/files/Mobiel_parkeren/Interface_Description_v7.6.pdf). A positive response means the car has valid parking rights in place, and the license plate scan data can be removed in 48 hours. For license plates with invalid parking rights, the case is transferred to the cities tax department, which connects to the RDW database to link the license plate with the car ownership data, and to deliver a fine.
System architecture image
- Non-discrimination Show More Show Less
Promotion and realisation of equality in the use of the service.
The automated control does not process or use information that would have been identified as potential sources of discrimination against people impacted by the automated process. The service works the same way for all license plates regardless of the car model, age, or the owner’s profile.
- References Show More Show Less
Legal basis description
Live service address
- Risk management Show More Show Less
Risks related to the system and its use and their management methods.
The system’s overall risk level is low. The key risk is that the system could incorrectly recognize a license plate and someone will be fined who does not deserve it.
This could happen if a character on the license plate is incorrectly recognized by both the algorithm and the inspector. To manage this risk, people are given the opportunity to object in writing via a website (naheffingsaanslag.amsterdam.nl) within 6 weeks. Anyone who objects will be given the opportunity to see the photo of the license plate and a situation photo, if available. Any bystanders, unrelated license plates and other privacy-sensitive information are made unrecognizable in those images.
Was this information useful?