In July 2021 APCOA entered into a strategic cooperation with Signatur ITS. The use of the ANPR MultiMotorAI SaaS enables APCOA to offer an unrivaled ANPR performance for parking customers and superior value to parking facility owners. APCOA is now forwarding some 150,000 images per day to the ANPR MultiMotorAI and the number is growing by the month.
APCOA has for more than 60 years offered parking solutions in Norway, and is today the market leader with 1,600+ facilities across the country. The company is part of APCOA PARKING GROUP, Europe's no. 1 car park operator and service provider with some 12,000 facilities in 13 countries.
APCOA is at the forefront when it comes to introducing innovations. ANPR is a good example. Today, a large part of the company’s parking transactions is handled by this technology: A vehicle's number plate is captured by cameras when entering and exiting the parking facility, parking time is calculated and the owner’s credit card is charged or an invoice is issued. All takes place in a fully automated and very user-friendly process.
APCOA's in-house ANPR system is state-of-the-art, but achieving 100 % read rate at all times is challenging. Dirt, snow, damages, etc. lower the readability of number plates, which in turn leads to loss of income, errors and the need for costly manual intervention. For APCOA with large transaction volumes, improving ANPR performance has substantial effect on income generation, cost reduction and the parking customer’s user experience and satisfaction.
The Norwegian Public Roads Administration (NPRA) is a government agency responsible for national and county roads in Norway. Signatur ITS has cooperated with the NPRA on various projects.
Presently we are working on a R&D project where the goal is to improve the method used for capturing average speed violations. Today, the national Automatic Speed Control system uses an ANPR-based method which compares the number plate readings from the entry and exit points of the control zone to find the same vehicle, in order to calculate its average speed.
Difficult weather conditions, no fixed illumination, etc. lowers average ANPR read rates. In addition, number plates could be very dirty, damaged or even manipulated. No or incomplete ANPR readings at entry and/or exit points mean that the average speed cannot be calculated and a violator escapes the consequences of speeding. This undermines the system's effect on increased traffic safety, and is also a problem with regard to just and fair enforcement.
In the R&D project Signatur ITS has developed and is testing a new method for matching images from entry and exit points. The method is based on AI technology, so-called Artificial Neural Networks (ANN). Not only the license plate but the whole front of the vehicle is used to match entry and exit images. This makes the method far less vulnerable to typical ANPR challenges as described above. Preliminary test results show a substantial improvement in matching rates.
Toll road company Fjellinjen is the main financial benefactor to road construction and public transport in the capital region of Norway with some € 350 million in yearly contribution. Electronic tags are the main method for collecting tolls, but automatic reading of number plates (ANPR) is also widely used. Some ¼ of passing through Fjellinjen's toll points is identified with images. The ANPR system's performance is therefore critical for Fjellinjen's ability to achieve high yields and cost efficient toll collection.
Large transaction volumes necessitate a high degree of automation, including automatic invoicing. However, a fully automated process puts a high demand on performance. Fjellinjen does not want to invoke customers' dissatisfaction and frustration, not to mention bad press and ridicule in the media, by sending invoices to the wrong addresses.
Fjellinjen’s ANPR system must not only have a high read rate, it must also give a high degree of certainty on results. Though ANPR software provides confidence numbers, this has proved to give false certainty. Consequently, Fjellinjen must send a substantial number of correct ANPR results to manual handling because they need to verify that results are correct.
This was the motivation for Fjellinjen to ask Signatur ITS to develop and test a new ANPR setup which could give the required certainty of results. The project led to the development of two very strong qualification concepts; Consistency and Consensus. With these two tools we were able to reduce the portion of ANPR results that had to be routed to manual verification by 250 %. The concepts have later on been refined and integrated into our ANPR MultiMotorAI solution.
BT Signaal was a pioneer in the development and operation of systems for electronic payment services, in particular within tolling. The company started tolling operations in 1953 and eventually became responsible for some half of Norway's 50+ tolling projects, including several ferry tolling projects.
"When the Norwegian tolling sector was reorganized into five regional tolling companies, BT Signaal was transformed and today makes up the main part of Ferde, the tolling company responsible for the southern and western parts of Norway.",
Signatur ITS cooperated with BT Signaal on a development project for a new ANPR based toll collection system. A legacy from this cooperation is that we focus on offering non-intrusive, plug-in modules (e.g. the ANPR MultiMotorAI) that boosts performance of customers' existing ANPR systems. As was the case for BT Signaal, most of our customers today don't need another standard ANPR solution but a service that can make good performance become excellent.