Automatic Number Plate Recognition (ANPR/ALPR) Software

INTRODUCTION TO AUTOMATIC NUMBER PLATE RECOGNITION

What Is An ANPR / LPR / ALPR Software?

ANPR (Automatic Number Plate Recognition), LPR (License Plate Recognition), or ALPR (Automatic License Plate Recognition): they all stand for an image processing software that is capable of recognizing license plates and transforming them into digital data.
ANPR or LPR? For further reading about the terminology, we suggest checking this blog article.

How ANPR Software Works

The operation of any ANPR system is comprised of three main elements.

1. Vehicle Detection & Image Capture

At the front end of any ANPR system, a camera captures images of the license plates. The camera is important as it needs to capture images that are optimal for ANPR purposes. This is crucial for the overall performance of the system. The best ANPR results are achieved by using cameras designed especially for ANPR.

Adaptive Recognition offers a wide range of such specialized ANPR cameras.

2. Image Preselection and ANPR

ANPR systems read the plate text and identify the plate type from a preselected set of captured images.

Automated recognition involves several steps like image normalization and enhancement, detecting the vehicle in the image, then finding the plate on the image. The final step is done by the OCR (Optical Character Recognition) algorithm that recognizes the individual characters and transforms them into digital text data.

For more information about image preselection and triggering, download our free e-book.

3. ANPR Data Recording and End-User Application

Besides the individual characters and the type of the vehicle plate, the CARMEN® ANPR also returns plenty of additional information, such as an image with the recognized plate(s), and the confidence level assigned to each character and a confidence level for the whole plate. The result also includes the nationality and the color of the plate (in selected regions/counties), with time-stamp and optionally geographical coordinates.

Once all license plate data have been saved to a database, the data record serves as input to the end user’s business logic. Automatic number plate recognition may be a key component of vehicle access control, traffic and toll enforcement and many other areas.

Requirements for Proper ANPR Results

ANPR accuracy depends on two major factors:

• the performance of the ANPR software itself;
• and the quality of input the ANPR software receives.

Higher quality input means a higher chance of 100% accuracy. In general, if an image is hard to recognize for humans, it will be similarly hard to recognize for an ANPR system. The license plate on the captured images should not be too distorted, blurred, or over-exposed. For this, we offer a complete guideline in our 24/7 ANPR image guide.

Application Areas Benefiting from ANPR

ANPR can be used in any road traffic environment from parking houses to highways, with a wide range of benefits.

Access control systems use it for a license plate-based entry/exit control. Modern highways use ANPR for automated toll control. Traffic management systems use license plate recognition for identifying wanted vehicles, detecting speeders and violators. ANPR is used for congestion charging, journey time measurement (JTMS), and more in smart cities.

For more information, head to our relevant industry pages: Tolling systems, Traffic safety, Access control systems.

Choosing the Right ANPR Software

When checking ANPR software options, the following factors should be considered:

• How many lanes and cameras the system will manage
• The frequency of vehicles – or in other words, how many vehicle number plates should be identified within an hour or a minute
• The type of license plates – in the given environment, what nationality of vehicles will be likely to occur (For instance, the expected nationalities are different in a border control system and in a city center)
• The preferred location of ANPR processing – on-premise and cloud systems are both available. Cloud architecture is rather new in the ANPR world. For more information, check out our ANPR Cloud service.

Depending on these factors, Carmen® offers ANPR software versions for 1-core, 2-core, and 4-core CPU processing and various engines for local, regional and even global geographics (on Cloud).

ANPR with Make and Model Recognition

With today’s contemporary deep learning technologies, ITS systems detect the category, weight, or license plate and the brand, model and color of the passing vehicles. Linking ANPR with make, model and color data is particularly important when detecting replaced or stolen license plates.

Learn more about make and model recognition.