Advanced number plate recognition


Following recent media reports surrounding the plans of local councils in the UK to replace speed cameras with advanced number plate recognition (ANPR) “average speed” cameras, it appears that the system could pose its own benefits in terms of safety and security.

ANPR systems capture images and use optical character recognition to identify number plates, commonly using six algorithms. So here is the technical bit. The first, plate localisation, is responsible for finding and isolating the number plate, while the second (orientation and sizing), third (normalisation) and fourth (character segmentation) improves and renders the image to identify individual characters. The fifth algorithm, optical character recognition, identifies the specific characters, then the sixth, syntactical/geometrical analysis, checks characters and positions against specific rules for the country of origin.

When carried out at a lane site, algorithms can be processed and stored – along with other information, such as date-time and lane identification – in just 250 milliseconds. The information is then converted into small data packets and transmitted to remote computers for further processing if necessary.

ANPR systems are widely deployed across US states, but have been created controversy elsewhere. In 2008, Germany’s Federal Constitutional Court found that ANPR systems violated the right to privacy in certain circumstances, and that its retention of information without a pre-determined use, such as tracking suspected terrorists, was in violation of German law.

However, with the implementation of such technology very much on the agenda of UK councils, Mirasys managing director Iain Cameron feels that there are a number of potential benefits.

What surveillance infrastructure would be needed to support the transition from speed cameras to the advanced number plate recognition technique?

Iain Cameron said: Aside from cabling, the majority of infrastructure is already in place. As the images of number plates are taken, the data from the average speed cameras would need to be forwarded to a centralised point, or a local processing server, in order to be further processed. This would mainly be done using cabling, so that would need to be established in order to support the transition. Aside from that, the majority of infrastructure is already in place.

What positive implications could the ANPR system potentially have on security?

IC: From a security perspective, the ability to track and determine the whereabouts of any particular vehicle could potentially be advantageous, as the information would be centralised and then made available to particular emergency services as and when needed.

If you take a bank robbery as an example, emergency services would be able to use the system to track any vehicle involved, and its movements, in order to find or narrow down the location of those responsible.

And for road safety?

IC: From a road safety perspective, the statistics tend to show that this system is more effective than regular speed cameras as it measures average speed over a set distance and time, rather than one particular measurement at a definite location. The speed cameras currently in use can be easily avoided by simply slowing down when they’re approached, with drivers speeding up again once they’ve been passed, which can pose its own road safety problems. They’re also easy to find and are marked on a number of devices, such as satellite navigation systems. Average speed cameras bypass this.

What complexities could the system potentially face?

IC: In terms of the ANPR, the storage and maintenance of the data itself would be relatively easy. Mirasys has considerable experience of dealing with number plate databases as well, so the level of expertise required is also in place already. Difficulties could arise when the system calculates the average speed over the distance and time taken, as a number of algorithms are used to do this. But the main complexity would be in transferring data to the necessary central processing server, which would require the infrastructure, such as cabling, to be up to scratch.


How would the system benefit crime prevention, with the implementation of the CCTV cloud concept?

IC: The cloud concept could create benefits for a number of agencies, as the data would be centralised on a single system. This would allow tighter regulation and increased control of the data, which can sometimes breach ITO guidelines in its present format. With the data centrally located, it is easier to manage as the cloud will take care of that.


Information obtained from