IndigoVision’s License Plate Recognition system is powered by InnoWare and has been fully integrated into Control Center. Here Giovanni Mariani President and CEO of InnoWare, looks at the history of automatic number plate reading systems and how it’s evolved in recent years.
A bit of history
Automatic number plate reading systems (ANPR) were born at the end of the 90s of the last century and, thanks to the type of algorithms used, they were a first (very embryonic) example of the application of artificial intelligence in the field of Intelligent Transportation Systems (ITS).
The first systems
Given the modest computational capabilities of the general purpose CPUs of that time (remember the Intel 486 at 33 MHz?) and the resolution of video sensors (typically CIF or SIF), the first ANPR systems able to read the license plates of moving vehicles were based on proprietary hardware and therefore extremely expensive.
There were cheaper solutions based on standard hardware, but they are not necessarily usable for access control.
Cost reduction and application diversification
Over time, thanks to the rapid and continuous evolution of computer technology, both hardware and software, it has been possible to use commercial processing systems and normal video surveillance cameras, with a consequent reduction in prices, thus favoring the spread of ANPR systems in an ever increasing number of fields.
Recent market research predicts that ANPR application sales will reach $ 800 million over the next 5 years, with an average annual growth of 16.4 percent.
The main applications of license plate reading systems include:
– access control
– parking management
– border control
– payment of tolls
– control of average speeds
– infringement control
This last type of application, thanks to the spread of geolocation technologies and the availability of centralized police force databases, is currently enjoying particular success.
In fact, thanks to tools like InnoWeb, police officers can easily monitor vehicle traffic through a web page, both from the Control Center and from mobile devices, with immediate notification of the infringing vehicles (speed, red light crossing, insurance and / or revision due, etc. ..) or belonging to black lists (stolen vehicles, reported for having been in criminal actions, etc.).
Evolution: ANPR on the EDGE
For a long time the architecture of the ANPR applications has followed that of the classic video surveillance systems.
In fact, in most cases, license plate recognition occurs through the use of software running on particularly powerful computers (back-end), analyzing the video streams coming from cameras distributed throughout the territory (front-end) sent to public or private monitoring centers.
As the resolution of the cameras increases, limitations in this type of architecture have become apparent, such as:
– high bandwidth demand
– increase in the power required to process frames consisting of millions of pixels.
The natural solution to this type of problem is the on-the-edge distribution of the intelligence necessary for the recognition of license plates, made possible by the miniaturization and increased power of the processing units.
Explore IndigoVision’s License Plate Recognition system powered by InnoWare by clicking the button below.