The Truth About License Plate Recognition Technology

After numerous questions about the merits and drawbacks of cameras designed to capture images of the license plates of vehicles entering and exiting a property, I wanted to put something together to clarify the two approaches to this application. These cameras are often referred to interchangeably as “license plate cameras” and “license plate readers”. In fact, these terms are not interchangeable and refer to two completely different types of systems, albeit with some functional overlap.

License plate cameras are video cameras that allow for real-time and recorded video of vehicles entering through the associated lanes. While model selection can vary depending on the application, these are “standard” cameras that, depending on lighting, weather conditions, and vehicle speed, etc., usually (but not always) provide the operator with the ability to view and recognize the image of a license plate and manually log the number. They also often capture information on the vehicle (make, model, color, condition) and sometimes the occupants, depending on placement and environmental conditions.

License plate readers are specialized cameras designed to capture tag numbers off of license plates. With the associated software (these are sold as systems), they attempt to automatically read and record a list of license plate numbers in an alphanumeric data format. They do not capture images of the vehicle, and are designed for this one specific purpose. They either provide this information in an Excel type spreadsheet format for manual comparison, or require integration into third-party databases that are only available to law enforcement agencies. When integrated, the systems are designed to “flag” a vehicle to an operator or via a printout when a tag matches one in the database. Together with the software, these are commonly referred to as license plate recognition systems.

license plate recognition system is far more costly than a camera capturing an image, and the database integration usually incurs recurring fees. These databases are not generally searchable, limiting the value of these systems for forensic examination of data to determine patterns and threats. They are not infallible, and rarely work as well once installed as the demos suggest. According to numerous sources, their primary purpose seems to be as a revenue generation system, allowing the automatic generation of lists that can be utilized by the local police jurisdiction to issue a traffic summons where appropriate.

There are also privacy concerns regarding such systems, particularly concerning the retention and disposition of the data collected by these systems. The American Civil Liberties Union (ACLU) is pursuing legal action on a number of fronts and the ultimate legality and usefulness of this equipment appears likely to be further defined and refined.

In summary, these two types of systems differ widely in functionality, usefulness, and cost (roughly ten to twenty times the cost for a standalone permanently mounted license plate recognition system versus a high-end high resolution IP camera).  If we are asked to specify them for a law enforcement agency affiliated with a client, we would require significant input from the agency to ensure that they understand the limitations of such systems and managing their expectations. While we continue to stay on top of this technology, I saw several of these systems in action at the security industry trade show ISC West 2014 in Las Vegas and my opinion has not changed. In fact, during one demo, the operator clicked on a license plate number that had been recognized and it brought up a video image of a plate that was not the same as the listed number. And that was based on demo footage, not real world conditions.

Some information and assessment of the operation of these systems, based on personal experience, was provided by Lieutenant Paul Sorrentino, Support Division Commander at the Township of Hamilton Police Department in New Jersey.