ANPR (Automatic Number Plate Recognition)

ANPR has been around for approximately a decade. Up until recently, it was used mainly by the police and military. The technology involved with the system has become far more reliable, and currently is capable of reading and correctly identifying a vehicle registration at up to 100 yards with the vehicle doing in excess of 100 mph.

The software we us in our system is the Talon software developed by NDI Recognition Systems primarily for the Police both in England and the US. It has proved to be extremely effective in both countries, so much so that it is still the preferred software in use. The software is tried, tested and proven over extended periods and has served the Police well in helping to apprehend criminals.

DMG operates a standalone product called Pay For Fuel If you are interested in how ANPR can be applied in a real life situation to both bring criminals to justice whilst saving money for the operator, visit our Pay For Fuel site.

The principal behind ANPR is very similar to face recognition and fingerprint recognition systems. The software engine finds the presence of a vehicle registration plate. It has a sophisticated 'plate finding' algorithm and this algorithm very accurately detects the presence of a licence plate in the camera's field of view. When a licence plate is detected, a trigger is activated which initiates a suite of software recognition algorithms. These algorithms identify the position of the plate in the image, extract the plate image, apply geometric distortion compensation if necessary and segment the plate into individual characters.

Each character is then fed into a software neural network. This network uses a unique mathematical perceptron model, unlike multi-layer perceptron models (MLP) the software model cannot suffer from 'over learning' and a reduction in accuracy. The network produces an estimate for every character input resulting in a string of characters, which is its best estimate for the number plate. For each passing vehicle multiple recognition results are obtained and our advanced post processor will analyse and produce the best result.

These features enable our system to be over 99.7% accurate on recognised plates.

The cameras used for ANPR systems are far more accurate than normal CCTV type cameras. They are designed specifically for the task of reading registration plates and will do so in day or night, sunny or cloudy situations.

DMG supply and install state of the art access control and barrier systems as part of our products and services portfolio. One aspect of this involves ANPR recognition. As with our forecourt systems, the registration plate is recognised and identified. If the plate is listed in the system database as ok, a signal is sent to the access control to activate the gate or barrier.

Classic applications of this system are private residences with sophisticated security systems or office staff car parks. It can also be used within private and public car parks to both grant access and to record arrivals and departures to determine if an overstay has occurred which could warrand an excess charge. Another application is within educational establishments where the system can differentiate between students and teachers and grant access to the correct car park.

The access control can also be used to log personnel on site and grant permissions to their pass cards.

VISCE – National Fixed Site Database

As part of our Pay for Fuel service, we are able to catalogue non paying offenders into the standalone VISCE database with live feeds from all fuel stations within the system. VISCE stores images and details of all readings taken from all garages. If a drive off occurs, the vehicle can be identified and the information passed to the police. The vehicle details are then downloaded to all the garages within the system. If the car attempts to enter any of the garages, an alarm will sound warning the attendant.