LONDON, July 21 — Automatic number-plate recognition is a technology that has been quietly gathering data on drivers for decades, yet most motorists still have little idea when or where their plates are being read. The mechanics here are straightforward: a camera captures an image, optical character recognition extracts the alphanumeric characters, and a database logs the vehicle’s location. The catch is that the same system that catches stolen cars and killer drivers also enables a permanent record of movement that critics call mass surveillance.
ANPR was invented in 1976 at the Police Scientific Development Branch in Britain. Prototype systems were operational by 1979, and contracts were awarded to produce industrial systems — first to EMI Electronics, then to Computer Recognition Systems, now part of Jenoptik, in Wokingham, UK.
Early trials were run on the A1 road and at the Dartford Tunnel. The first arrest through detection of a stolen car came in 1981. On closer reading, that means the technology existed for a quarter of a century before it was used to solve a murder: the first documented case came in November 2005, in Bradford, UK, where ANPR played a vital role in locating and convicting the killers of Sharon Beshenivsky.
But the spread of ANPR was not immediate. The technology did not become widely used until cheaper and easier-to-use software was developed during the 1990s.
The collection of ANPR data for future use — storing location records to solve then-unidentified crimes — was documented in the early 2000s. The software runs on standard home computer hardware and can be linked to other applications or databases, meaning the barrier to entry is lower than the hardware costs suggest.
How the system works
The software side uses a series of image manipulation techniques to detect, normalise and enhance the image of the number plate, then applies optical character recognition to extract the alphanumerics. The system can work in one of two basic ways. The first approach performs the entire process at the lane location in real-time, completing the capture of plate alphanumeric, date-time, lane identification, and any other required information.
The second approach transmits all images from many lanes to a remote computer location and performs the OCR there at a later point. Cameras often use infrared lighting to allow captures at any time of day or night.
The technology must account for plate variations from place to place, a significant technical hurdle given the differences in design, font, and layout across jurisdictions. ANPR can also store the images themselves, not just the text, and some systems are configurable to store a photograph of the driver.
Uses beyond law enforcement
Police forces around the world use ANPR to check if a vehicle is registered or licensed. But the technology has commercial and administrative applications as well. Electronic toll collection on pay-per-use roads relies on it, and highways agencies use it to catalogue the movements of traffic.
The claim is that this is efficiency; the catch is that each use case adds another layer of permanent data collection. Privacy issues have caused concerns about ANPR, including government tracking of citizens’ movements, misidentification, high error rates, and increased government spending.
Critics have described it as a form of mass surveillance. The technology is also known by several names — automatic license-plate recognition, automatic vehicle identification, license-plate recognition, and vehicle license-plate recognition — a proliferation of names that does nothing to make it less opaque to the public. What to watch next: as the cost of hardware continues to drop and the processing power of standard computers rises, the spread of ANPR is likely to accelerate.
The London congestion charge project already demonstrates the server-farm model at scale. The question is whether privacy safeguards will keep pace with technical capability — or whether the permanent logging of every vehicle’s location will become the default, not the exception.
