Automatic number plate recognition or ANPR was first developed in 1976 in the UK. By 1979, there were several prototypes in use. Early systems were used in several places in the UK. The first arrest stemming from ANPR took place in 1981. The technology continued developing but was not widely used until the 1990s. Data collection for future use became more common in the early 2000s. In 2005, ANPR was used to help solve a murder case, and it was used soon after to track down several murder suspects.
How ANPR Works
The technology uses several image manipulations to recognize and decode the information on a license plate. It works with a software that runs on a standard computer, and it can be linked to multiple databases. When the alphanumeric characters are extracted, the system is deployed. It starts by performing the recognition process at the location where a plate is being read. After this, it sends data and images from multiple lanes on the road to a computer for the remainder of the recognition process. However, the process may not be completed until a later time in some instances. When information is decoded at the site instead, it takes only a fraction of a second to complete. The information is still sent to another computer if it is instantly decoded, and it may be further analyzed if necessary.
In some countries, the font on plates may be changed to enhance recognition. This usually involves making small gaps between letters or numbers that look similar. Optical character recognition is more accurate with such measures in place. With some plates, there are variations in font sizing and positioning. An ANPR system must be designed to compensate for such differences. Each country has different systems based on its arrangement of license plate numbers and letters. When cameras capture the data, they may be connected to closed-circuit systems and used for internal purposes. For example, a financial company with a secured entry that tracks everyone who comes into a parking garage would use such a system. If a crime happened, security personnel could easily retrieve license plate numbers. There may also be live road-rule enforcement camera systems, which are used by law enforcement agencies.
There is a set of algorithms included in plate recognition. They include:
- Plate localization to distinguish the plate from surrounding details.
- Plate sizing and orientation to compensate for a crooked alignment of the plate.
- Normalization to enhance the contrast and brightness of the image.
- Character segmentation to distinguish individual numbers and letters on the plate.
- Optical character recognition to confirm characters.
- Geometrical or syntactical analysis to match characters with alignment rules in the country.
- Value averaging to recognize value using several fields and images for enhanced accuracy.
Plate recognition is often used today to track moving violations in traffic and automatically send out citations. And these cameras are becoming more and common so watch those yellow lights.