The CARMEN® Freeflow License Plate Recognition Engine possesses all the features a high quality license plate recognition system may require: it is highly accurate, fail safe and stable, fast and intelligent.

Meanwhile it provides the highest technological performance amongst today's license plate readers, CARMEN® FreeFlow offers great flexibility for manufacturers and integrators of car plate recognition systems: it can be integrated into practically any automatic number plate recognition system which needs automatic vehicle recognition.

There is no compromise. The CARMEN® FreeFlow's Software Development Kit (SDK) makes integration of optical plate recognition extremely easy, fast and cost effective. And the resulted image processing system will have the highest performance - both in terms of recognition accuracy and speed - of optical license plate recognition.

The CARMEN® Freeflow offers general-purpose license plate reading from digital pictures of any tpye of sources. Whether it is a megapixel-size, 12bits/pixel high dynamic range, high resolution digital image or a 384x288 size colour CCTV video picture provided as input, (either from file or memory) CARMEN® provides ballanced reading of car plates.

While there is a vast space for different configurations - you can freely set the working parameters of CARMEN® to best fit your actual system's requirements - you do not need to become a master of tuning parameters: CARMEN® is intelligent enough to provide high quality car plate recognition for very different type of images and plates with even the default settings.

There are automatic number plate recognition systems powered by CARMEN® with over 98.5% recognition rate (correct reading/total number of input) working 24h/day, 7days/week.

CARMEN® is a proven product - and a living automatic vehicle recognition technology.

The CARMEN® Freeflow ANPR Engine and its SDK is a PC based solution supporting today's most PC operating systems.

Supported Operating Systems	Windows Vista 32bit, XP, 2003, 2000 or Linux
Programming Languages (under Windows)	Visual C/C++ 6.0 Visual Basic 6.0 Borland Delphi Visual Basic .NET C# Java
Programming Languages (under Linux)	C/C++
SDK for easy integration	DLLs, ActiveX components, OCX files
Type of Plates	Recognition is country and font independent. Any Latin, Arabic, Chinese, Korean and Cyrillic characters can be recognized. State/country region recognition (optional)
Image Input	Still image from memory or file (BMP, JPEG,JPEG2000) Live video input (PAL or NTSC)
Trigger	Trigger is not necessary but recommended when recognizing from live video
Sample Processing Time	Standard engine: 50ms @ CPU 2GHz B&W images, 768x288 pixels (PAL/2) Quick engine: 7ms @ CPu 2GHz 768x288 pixels (PAL/2)
Processing Time Dependence	Image quality (complexity, noise level etc.) Image size Processing power (CPU speed) Parameter settings
Output	Plate number in ASCII/UNICODE Position of plate Positions of characters Tip list for each characters Confidence levels for each tips Colour of plate (optional) Country ID (optional) Locating more plates on one image
Documentation	Reference Manual in electronic format
System Requirements	Intel PIII. 1 GHz or higher CPU (128MB RAM) Free PCI / PC104+ slot or USB 2.0 port

Applications of CARMEN® FreeFlow While CARMEN® Freeflow is suitable for any automatic license plate recognition system, it is mainly used in Intelligent Transportation Systems (ITS), security systems and low enforcement systems. The main applications are: motorway toll collection (highway toll collection), traffic analysis, weight in motion, police law enforcement, state border control, enforcement of traffic regulations, vehicle theft prevention, assests protection, automation and simplification of airport, harbour and warehouse logistics, security monitoring of roads and checkpoints, vehicle surveillance, etc... As car plate recognition is the most natural way of automatic car identification, car license plate recognition can be used anywhere, where automatic, computerised vehicle identification is needed.