Automatic Number Plate Recognition: Unlocking the Potential of Computer Vision Technology

By Fouad Sabry

What is Automatic Number Plate Recognition

Automatic number-plate recognition is a technology that uses optical character recognition on images to read vehicle registration plates to create vehicle location data. It can use existing closed-circuit television, road-rule enforcement cameras, or cameras specifically designed for the task. ANPR is used by police forces around the world for law enforcement purposes, including checking if a vehicle is registered or licensed. It is also used for electronic toll collection on pay-per-use roads and as a method of cataloguing the movements of traffic, for example by highways agencies.

How you will benefit

(I) Insights, and validations about the following topics:

Chapter 1: Automatic number-plate recognition

Chapter 2: Intelligent transportation system

Chapter 3: Traffic enforcement camera

Chapter 4: Electronic toll collection

Chapter 5: Police car

Chapter 6: Open road tolling

Chapter 7: SPECS (speed camera)

Chapter 8: Road speed limit enforcement in Australia

Chapter 9: Roads Policing Unit

Chapter 10: Video tolling

(II) Answering the public top questions about automatic number plate recognition.

(III) Real world examples for the usage of automatic number plate recognition in many fields.

Who this book is for

Professionals, undergraduate and graduate students, enthusiasts, hobbyists, and those who want to go beyond basic knowledge or information for any kind of Automatic Number Plate Recognition.

• Language: English
• Publisher: One Billion Knowledgeable
• Release date: May 14, 2024

Book preview

Chapter 1: Automatic number-plate recognition

The process of reading car registration plates through photographs using optical character recognition to provide vehicle position data is known as automatic number-plate recognition (ANPR; see also various names below). It may make use of preexisting CCTV, road rule enforcement cameras, or purpose-built cameras. Police departments all across the globe utilize ANPR to conduct legal inspections, such as verifying a vehicle's registration and insurance. Highway departments, for instance, utilize it to keep track of traffic patterns and collect electronic tolls on pay-per-use routes.

Images acquired by the cameras and the text on the license plate may be saved using automatic number-plate recognition, and some systems can even be set up to save a snapshot of the driver. The camera can capture a picture at any time of day or night since infrared illumination is used in most systems. ANPR system must adjust for regional differences in license plate design.

Misidentification, high mistake rates, and greater government expense are just some of the privacy concerns raised by ANPR. Some have claimed that it constitutes a massive monitoring program.

Alternative Names for License Plate Readers:

Recognizing vehicles by their license plates mechanically (ALPR)

Motorized or computerized license plate reader (ALPR)

Identifying cars automatically (AVI)

Automatically generated number placement (ANPG)

License plate readers (CPR)

Plate reader technology (LPR)

Automatic license plate reading system (LAPI)

License plate reader on wheels (MLPR)

Automatic License Plate Reading Systems (VLPR)

Automatic License Plate Reader (VRI)

In 1976, the Police Scientific Development Branch in Britain developed the technology that would become ANPR.

The system's software is compatible with common desktop PC hardware and may be integrated with third-party programs and data stores. Optical character recognition (OCR) is used to extract the license plate's alphanumerics after a series of image processing methods are used to identify, normalize, and improve the picture of the number plate. There are two primary methods for deploying ANPR systems: the first allows for the entire process to be carried out at the lane location in real time, while the second involves transmitting images from multiple lanes to a remote computer location, which then carries out the OCR process at a later time. It takes around 250 ms to record the alphanumeric characters of a license plate, the time and date, the lane number, and any other necessary data at the lane location. This data may be kept locally at the lane, transferred to a distant computer for processing, or both. The second setup, a server farm, is used to manage intensive tasks like those involved in the London congestion charge system. More bandwidth-intensive transmission medium may be needed in such systems due to the frequent need to upload photographs to a distant server.

Camera pictures are processed using optical character recognition (OCR) software for use in ANPR. Small gaps were included in certain letters (such as P and R) when the Dutch transitioned to a new form of car registration plates in 2002, making them more unique and, hence, more readable to such systems. To be genuinely successful, ANPR systems need to be able to deal with the fact that certain license plate layouts employ different text sizes and placement. Even while many programs are adapted on a country-by-country basis, more complex systems are better able to handle worldwide versions.

Mobile units, often mounted on cars, may also be employed, together with fixed cameras already in place for road rule enforcement or closed-circuit television. To get a better look at the plates, some setups employ infrared cameras.

Technology advancements in the 1990s transformed automated number-plate recognition (ANPR) systems from cumbersome, costly, and stationary to portable, point and shoot devices. The development of software that could function on standard personal computers removed the necessity for precise knowledge of the angles, directions, sizes, and speeds at which the plates would cross the camera's field of view. More compact and affordable parts prompted a surge in adoption by police departments throughout the world. Officers may now patrol everyday with the advantage of license plate reading in real time, when they can intercept quickly thanks to the miniaturization of both the cameras and the computers that make this possible.

However useful they may be, mobile ANPRs are not without significant drawbacks. In the event of approaching traffic, one of the main challenges is making sure the computer and cameras can keep up with relative speeds of more than 100 mph (160 km/h). Since the equipment will be running off of the vehicle's electrical supply, it must be particularly space-efficient.

The camera's ability to read a license plate is impacted by a number of factors, one of which being the vehicle's relative speed. Time of day, weather, and camera angles may all impact an ANPR's capacity to generate an accurate scan, therefore algorithms must be able to account for these factors. Under these circumstances, the resolution and accuracy of a read may also be directly affected by the wavelengths of light used in the system.

When deciding where on a police vehicle to mount an ANPR camera, it is important to take into account its proximity to the license plates it will be scanning. Given the wide variety of missions and conditions, it might be difficult to determine how many cameras to use and where to place them for optimum results. Forward-facing, multi-lane, license-plate-reading cameras are essential for highway patrol. Cameras with a shorter range and smaller focus length are necessary for city police to capture plates on parked automobiles. Parking garages with rows of vehicles aligned perpendicular to one another need for cameras with very low focus lengths. The most cutting-edge systems are malleable, allowing for the installation of anywhere from one to four interchangeable cameras. Since a forward-facing camera is rendered useless in the presence of approaching traffic, states that only need rear license plates have an extra difficulty. In such a scenario, one camera may be inverted.

The program needs seven core algorithms in order to detect a license plate:

Finding and separating the plate in the image (plate localisation)

Plate orientation and sizing - corrects plate tilt and modifies measurements to meet specifications

Image brightness and contrast are normalized during the normalization process.

Separating out the characters on the plates is called character segmentation.

Optical character recognition

Analyzing characters and their placement in relation to regional grammar norms

Recognization value averaging over numerous fields/images to increase confidence in the final result, since every particular picture may suffer from blurring due to light flares, partial obscuration, or other factors.

The precision of the system is a function of the difficulty of these individual parts of the code. Some systems utilize edge detection methods during the third phase (normalization) to improve the visual contrast between the letters and the plate background. To further minimize distracting visual elements, a median filter may be used.

The program has to be resilient enough to handle a variety of problems. Among them are:

The plate's distance from the camera, or a poor-quality camera, both contribute to low file resolution.

hazy photos, especially those with motion blur

Overexposure, reflection, or shadows cause dim illumination and poor contrast.

A tow bar or other obstruction, or dirt on the plate, might make it difficult to read.

License plates on towed trailers, campers, and other vehicles sometimes have separate front and rear plates, so it's important to read both.

License plate reading interrupted by vehicle lane change

Typeface variant; often used for personalized license plates (some countries do not allow such plates, eliminating the problem)

Circumvention techniques

Disjointed action on the part of several nations. It is possible for two vehicles to have a plate number yet to have visually distinct plates if they are from separate nations or states.

Some of these issues can be fixed in the software, but generally it is up to the hardware designers to figure