Kapsch TrafficCom recently released a significant Automatic Number Plate Recognition (ANPR) software update. The update automatically recognizes license plates (also known as number plates), depending on the application.
The latest release features a reworked architecture with various deep-learning elements, resulting in a significant performance boost, according to Kapsch.
With the new ANPR software, artificial intelligence software was trained to accurately and reliably identify license plates with hundreds of thousands of images in a GDPR-compliant manner.
What is ANPR?
ANPR is a technology for reading and interpreting vehicle registration plates. It is indispensable for many modern tolling and traffic management applications, for example, to determine the correct toll rate for a vehicle in a barrier-free tolling system or to determine access rights for low-emission zones.
But, accurately identifying them is a challenge because many different types of license plates vary in legibility depending on:
- Weather conditions.
Julia Azfar, mobility expert at Kapsch TrafficCom, comments: “The higher the level of automation, the more efficient the system can be, as this reduces the need for costly manual checking.”
According to Azfar, continuous improvement is also a focus: "Especially in projects with long durations and high requirements, it can be worthwhile to benefit from the rapid technical development through regular updates. That’s why we work closely with partner companies and suppliers to combine the latest technologies on the market with our image processing and artificial intelligence expertise. That way, we can guarantee premium quality, innovation, and state-of-the-art services for our clients.”
Training AI Software
The software was "trained" with the help of the “Responsible Annotation” project.
To identify images, the software has to be trained with information about the image content in addition to just the plain images, for example, whether there is an Austrian or Italian license plate on a photo. Kapsch also operates in North American markets, as well.
This information is called annotation, and it is essential for the correct processing of the images by the system. If the software is fed with enough annotated images, it can subsequently process non-annotated images independently.
Kapsch TrafficCom carried out this annotation of images as part of a pilot project in Vienna that introduces disadvantaged people in the job market to new occupational fields. The annotation and validation of data is a new field of work that will grow strongly in the coming years due to the increasing use of AI. The Responsible Annotation project gives people at risk of exclusion a realistic pathway into the primary labor market.