Top ten things to look for in License Plate Recognition (ANPR)

Top ten things to look for in License Plate Recognition (ANPR)

In the security category automatic vehicle identification a lot of activity involves license plate recognition (LPR or ANPR). The field seems to progressing rapidly and is continuing to gain popularity. It seems that there is more and more awareness of the importance of the vehicle as a carrier in secured in environments that needs to identified accurately and swiftly. RFID based solutions, like the ones that rely on UHF readers and tags, are also popular, but the big advantage of Automatic Number Plate Recognition (ANPR) is that there is no need to install additional tags in or on the vehicle. Besides the security industry, also the traffic/mobility and parking industry have widely adopted license plate recognition technology.

But what should you consider when searching for an ANPR solution? What are the top ten things to look for in license plate recognition?

Top 10 things to look for in license plate recognition

10. Topology

The software that is used to convert an image to a string representing a license plate, can be installed on a server (in the cloud or on-premise) or it can be embedded on the camera device. The disadvantage of server based ANPR is that it does not offer standalone operation when connections get lost, that it is a more complicated topology with additional installation effort and that it may require additional licensing. The advantage is that a variety of suitable, but more or less standard, video cameras may be used and that the computing power of the server infrastructure may improve the performance of the system.

9. Robustness

ANPR cameras are used to identify vehicles and are therefor by definition mostly installed in outdoor situations. Sometimes the weather and environmental conditions may be harsh and challenging for the device. These products are installed in moist and hot, dry and sunny and cold and snowy conditions. Mechanically, technically and performance wise the cameras need to able to perform at adequate levels. Not only a few days but multiple years. Preferably without a great deal of maintenance. Check the experience vendors have with different circumstances and verify their willingness to extend the warranty period. It will tell you something about the experience they had with the enduring capability of their products.

8. Regional support

Probably the most challenging aspect of developing an ANPR camera is the great variety in license plates across all countries and in some cases even within the country. Colors, fonts, lay-out, reflectiveness and contrast in foreground and background colors vary enormously across license plates in use today. Every state in the USA and every country in Europe is using different colors, fonts and lay-out. Sometimes even within the country or state a variety of license plate schemes is used. Always verify with your vendor of choice if the license plates of countries and states, that are relevant to your installation, are supported simultaneously.

7. Technical design

An ANPR camera is different from normal surveillance cameras. Most ANPR-cameras today use IR-light to read the plate or white light if colors need to be identified. The focus of the camera is on the plate and not necessarily on the environment. The shutter in the camera must be optimized to deliver great quality freezed images. Zoom-lenses in cameras sometime are expected to degrade in performance over time. The best video surveillance camera may not be a good ANPR camera. And ANPR-cameras are usually not the best surveillance cameras. We would advise you to check with your supplier how the technical design of the camera is optimizing its performance.

6. Classification and incident detection (AI)

In a recent video we explained that modern ANPR cameras often are optionally equipped with an additional (surveillance) camera with AI capabilities. This additional camera can not only be used for traditional streaming of surveillance footage to a specific video management system, it could also enable things like classification of vehicles or the detection of security and safety related incidents. We all understand the enormous potential of video analytics, so it makes absolute sense if the vendor of your choice has a clear vision on how they will incorporate AI capabilities into their portfolio.

5. Mechanical design

Besides robustness (9), there are additional relevant requirements for ANPR cameras. Cameras are installed on gantries, poles, walls and sometimes inside stainless steel constructions. Is the size and shape of the camera and its mounting tools supporting the installation you have in mind? Is the cabling easily achieved and can you get easy access to the part of the camera that is used for configuration? Does the mechanical design consider solar influences or rainy conditions throughout the day?

4. Interfacing

Sometime ANPR cameras need to operate standalone. In that case the ANPR camera must support a white list and have a relay output to initiate the barrier. But in most cases the camera is connected to third party systems. Does the camera support all interfaces that are required? Do you need it to support IP, Wiegand, Serial or specifically OSDP connectivity? If there is an overview camera built in, does it support ONVIF or native support for VMS-systems from vendors like Milestone or Genetec? Is there additionally a well-documented API available to support enhanced interfacing?

3. Configurability

Every ANPR camera that is installed will require some level of configuration. You may want to check the configurability of your desired ANPR camera before procurement. Can you specify the area in the camera image that is relevant? Does the camera support multiple license plate detection and identification modes (like free-flow or triggered modes)? Can firmware versions be updated remotely? Is it possible to trigger events? Can you choose where to store images related to identification events? We would always advise to request a demo of the configuration options of the camera. It may seem dull, but it will give you insight into the capability of the camera and also into the real experience of the manufacturer.

2. Accuracy and performance

Accuracy and performance numbers are easily given. It is astonishing to see new vendors claiming an accuracy level in detection and identification of 100%. That may be possible in theory and in lab environments, but in real-life situations cameras have to deal with a variety of plates that are sometimes dirty or damages. License plates may be installed on a variety of vehicles. Accuracy and performance levels are influenced by circumstances and the quality of installation and commissioning. We would advise not to just look at a number, but ask for specified performance levels in situations that are comparable to your planned installation. And when in doubt, run a pilot.

1. R&D Capacity

License plates change constantly. New schemes, new fonts new colors: changes that will affect the performance and that need to be incorporated into new firmware versions. Your requirements may change. And the technical possibilities are rapidly changing as well. In the current security industry a lot of products are sold under OEM-agreements. We would advise to check the release planning and relevant R&D capacity of your potential vendor. It will tell you whether your investment is sufficiently future-proof.

We would advise you to speak to your supplier about these topics with your potential supplier. If your are looking for a great ANPR specialist, please have a look at the ANPR category of our online security trade show.

Tagged , , , ,

1 thought on “Top ten things to look for in License Plate Recognition (ANPR)

  1. […] because of agile working concepts. We also looked at Automatic Number Plate Recognition and made a top ten list of things to look for in ANPR. Another popular theme nowadays is cyber security. That is why we presented our top ten reasons why […]

Comments are closed.