All Collections
Methodology & Requirements
Automatic Number Plate Recognition (ANPR, LPR)
Automatic Number Plate Recognition (ANPR, LPR)

What is ANPR or LPR, its typical usage and benefits

Lukas Hruby avatar
Written by Lukas Hruby
Updated over a week ago

ANPR (Automatic Number Plate Recognition) is a technology that extends the standard cameras and video surveillance systems with the ability to read number plates on the vehicles automatically via specialized software. Based on the area, this functionality may also be called LPR (License Plate Recognition) or ALPR (Automatic License Plate Recognition). The meaning of these terms is the same.

ANPR technology has a huge impact on the transport industry. A license plate is a unique identifier of the vehicle and its recognition can serve many different purposes, e.g.:

  • O-D surveys

  • Transport macro modelling

  • Average travel time surveys

  • Advanced vehicle classification

  • Tolling

  • Speed measurement

In recent years, ANPR has often been a subject for privacy-related and mass-surveillance discussions. Vendors and providers of this technology must, therefore, secure the correct data-handling and avoid possible data leaks. However, in the right hands, ANPR is an irreplaceable source of information and must-have technology for every macro traffic modelling tool.

In GoodVision, we are using state-of-the-art ANPR technology that reaches top accuracy. Our optimized ANPR solution is able to provide both fingerprinted and plain text values. By using ANPR, GoodVision Video Insights platform users are able to enrich the current traffic data with these unique identifiers of vehicles, provided the processing requirements are fulfilled.

The license plate is a unique identifier of vehicles that is relatively small and usually placed low above the ground. For the system to be able to read number plates on the vehicles automatically, the recording has to meet some quite strict requirements. The basic, general rule is that if a human is able to read the license plate from the screen, the ANPR mechanism will most likely be able to read it too. And the other way round, if a human has a hard time reading the license plate, can read only part of it or needs to guess the shapes, our system won’t be able to read it correctly either. Let's check some more detailed requirements.

ANPR functionality in GoodVision Video Insights

ANPR feature in GoodVision Video Insights

GoodVision offers the possibility to combine its cutting edge traffic data extraction algorithms with the latest ANPR technology within a single platform. ANPR is a special type of report - basically a vehicle list which includes the information about license plates of the vehicles crossing a selected filter, whenever possible to “read” them by the system.

To gain information about license plates on an appropriate movement, visit the “Traffic exports” page of your Data source and choose the ANPR report. When choosing the right movement (filter), you’ll see the information about the proportion of suitable traffic objects that can be used for the ANPR report. The exact number of objects is visible also in the row “Sum of vehicles”. The processing cost is the price for the ANPR recognition.

Creating an ANPR report

Important thing to know is that the amount of suitable objects does not necessarily mean the number of recognized license plate numbers you’ll receive in the report. Suitable object is simply a vehicle that fulfills the prerequisites for the ANPR analysis - especially the pixel size on the scene which is set to 350 pixels in GVVI. If a vehicle is suitable for the ANPR, it has the minimum size of 350 pixels on the scene which enables the license plate reading. So it represents the number of vehicles that will enter the ANPR process. But there are other characteristics like camera angle, front/back of the vehicle heading towards the camera etc. that need to comply with the requirements.

Remember that not all videos are suitable for ANPR processing and the requirements here are pretty strict. The camera angle, distance, resolution and frame rate all play their role in the ability of the system to recognize and track the uniquely identified objects. The results will be more or less precise depending on the video quality and the quality of camera placement. Check out our more detailed guidelines and requirements in the following article.

It is advisable to pay attention to the scene description when intending to have license plate information. Although all objects from the chosen movement trajectory (bigger than the pixel limit) are included in the ANPR analysis, it might be better to place a particular line just for the ANPR report to a spot where the license plate is expected to be the most readable than to use just the common traffic filters.

OpenAPI integrations (UK DVLA, DOT, etc)

Using OpenData from the governmental databases and licence/number plates, users are able to obtain even more detailed information about the traffic in the monitored areas. GoodVision is cooperating with all state agencies that provide open data about vehicles. A typical example of the acquired data from the Driver and Vehicle Licencing Agency (United Kingdom) is:

  • Colour

  • CO2 emissions and Euro emission classification

  • Model/make and other technical data about the vehicle (engine capacity, weight, etc)

  • The month of the first registration

  • Commercial / private vehicle classification

  • … and many others

Did this answer your question?