Frederikssund, Denmark

99.9% Recognition Accuracy for an Automated Tolling Project in Denmark

Thanks to Adaptive Recognition’ Carmen® ANPR API, road users registered on a website with their number plates and card details get automatically charged when crossing Crown Princess Mary’s Bridge.

About the client

Fjord Connection Frederikssund

Fjord Connection Frederikssund is owned by the Ministry of Transport of Denmark. It is an independent public company that handles the 10 km long, 4-lane motorway south of the town of Frederikssund.

About the Project

Crossing the Roskilde Fjord at Frederikssund had for a long time been a tiresome experience. This was due to heavy traffic and long queues building up at Crown Prince Frederik’s Bridge. To relieve the pressure on this single passage, a new bridge and expressway was built to connect the two sides of the fjord. This section of the highway is now enhanced with an automated tolling system. This way, vehicles do not need to stop for toll payments.

Crown Princess Mary’s Bridge was inaugurated in the autumn of 2019. The 1.4 km long bridge is part of Fjordlandsvej, a 10 km long expressway. Heavy lorries now use this alternative route, further away from the city of Frederikssund to cross the fjord via Crown Princess Mary’s Bridge. In fact, the Construction Act governing the new fjord crossing stipulates that lorries with or without trailers and vehicles over 3,500 kg must use the new bridge.

Other road users can choose between Crown Princess Mary’s Bridge and Crown Prince Frederik’s Bridge for crossing the fjord.

The construction of Crown Princess Mary’s Bridge over Roskilde fjord

Road Usage Data

In 2020 an average of 3,284 vehicles crossed the bridge in the course of a day. 82% of these were percent passenger cars, 13.6% trucks, and 4.4% motorcycles.

Different toll charges apply for these different vehicle categories.

Automated Tolling Based on License Plate Data

Road users can register at fjordpay.dk website with their license plates to sign up for an automatic payment scheme. After registration, they no longer need to worry about paying for crossing Crown Princess Mary’s Bridge. The system reads vehicle license plates each time they cross the bridge. At the same time, it automatically deducts the fee from the payment card they specified upon registration.

Registration is free, and they only need to pay for the actual crossings.

Additionally, limited-period agreements exist for visitors and single-time road users. In 2020, nearly 50,000 users registered with this service.

A Freeflow Traffic Experience for All Road Users 

As everything happens automatically, stopping and queuing at toll plazas is no longer a problem.

The system developed by our partner, Kapsch TrafficCom, also automatically registers the license plates of vehicles that cross the bridge without having an agreement. In this case, the owner of the vehicle receives an invoice by post after the passage.

How the System Works 

The Fjord Connection toll plaza uses Adaptive Recognition’ Carmen® FreeFlow. This software integrates into any system, with optional make and model recognition (MMR) and even hazardous material code (ADR and IMO) recognition. The software works with 3rd-party cameras and sensors.

In the case of the automated tolling system for Crown Princess Mary’s Bridge, Kapsch TrafficCom used advanced VDX sensors to distinguish between different vehicle types.

Despite the high speed of the cars on the motorway, the sensor can take 3D images of several lanes simultaneously.

Carmen is powerful enough to allow for real-time detection and profiling of vehicles according to their height, width, axle counts, etc.  

Efficiency: Let the Numbers Do the Talking

According to an audit carried out by Sund & Bælt, a consultancy firm specializing in infrastructure development projects, the passage detection ratio is 99.83 percent, of which number plate recognition is 99.99 percent accurate. 

A final word from our partner:

Adaptive Recognition has been a supportive and responsive partner to obtain the highest possible automatic license plate recognition of the vehicles.

Mikael Hejel, Kapsch TrafficCom