Artificial intelligence can read car number plates

When solving crimes, even renowned forensics experts sometimes reach their limits, for example when the quality of video material from near the scene of a crime is very poor. A research project hopes to remedy this situation: In conjunction with the Bundeskriminalamt (BKA) (Federal Criminal Police Office) and IT company Secunet, FAU will be working on a technological solution during the next two years that will allow car number plates to be read even on blurred images. The project is being funded by the Federal Ministry of Education and Research as part of the ‘Research for Civil Security’ programme.

FAU is contributing its expertise in artificial intelligence (AI) and signal processing to the project. Dr. Christian Riess from the Chair of Computer Science 1 IT Security Infrastructures, and Dr. Jürgen Seiler from the Chair of Multimedia Communications and Signal Processing are leading the project.

Images and videos that contain car number plates can provide investigative authorities with valuable leads. ‘It’s often very difficult to identify anything at all, either because of motion blurring, unfavourable lighting or suboptimal camera technology,’ says Anatol Maier, who is currently a doctoral candidate at the Chair of Computer Science 1.

In many cases, it is virtually impossible to clearly identify car number plates with the naked eye or by using conventional technical methods of analysis, meaning that investigations have to be discontinued. It is hoped that this new approach using AI should make a difference and make police work more efficient. ‘The improved signal processing and analysis using AI will enable us in future to analyse evidence that cannot be used or is overlooked entirely with today’s technology,’ says Anatol Maier.

Website of the Chairs involved

/FAU Public Release. The material in this public release comes from the originating organization and may be edited for clarity, style and length. View in full here.

Leave a Reply

Your email address will not be published. Required fields are marked *