|
|
Ultra High Recognition Accuracy
Intelligent OCR Processing
Compliancy With Standards
Flexible Opened Interfaces
Stable Operation
Robust Design
High Durability
Easy Operation
Standard Interfaces
High Functionality
|
License Plate Recognition / Number Plate Recognition
General Information
Overview
License plate recognition (or as frequently called 'number plate recognition')
is a special form of optical character recognition (OCR).
Actually, today optical character recognition (OCR) technology is considered a type of
tehcnology - mainly software - that lets you scan paper documents and turn them into electronic,
editable files. License plate recognition has somewhat similar function: it should scan images
of vehicles and convert the picture of the registration number into ASCII text.
However, license plate recognition is far more than OCR.
There are several differences between paper documents and photos taken from vehicles. Some of these are:
·scanned documents usually have 300DPI or higher resolution: vehicle photos rarely reach 300DPI
·scanned documents usually have 1bit/pixel: vehicle photos have at least 8bits/pixel
·scanned documents usually have hundreds of letters, organised in rows and coloumts: an averge plate number contains 6-8 characters
·scanned documents are two-dimensional images: vehicles on photos are in real three-dimensions
·documents usually have text written in a living language: vehicle license plates are raw data
As a result, both traditional and sophisticated OCR system fail badly when trying to read vehicle license plates.
As plate recognition systems require special hi-tech solutions both in software and hardware technologies,
there are only a few viable systems exist, all involve the results of latest researches of computer sciences.
Someone may ask, "OK, I know OCR, but why plate numbers should be read at all?"
The vehicle's primary identification sign is its license number (or the registration number) written well visible
on the license plate or number plate of the vehicle. Therefore reading the license plate of the vehicle equals to
learning the registration number of the vehicle, and knowing the registration number means having the vehicle identified.
Any automated system that is able to read the plate number of a vehicle is actually able to automatically identify that vehicle.
License plate recognition have undoubtable advantages over other identification concepts:
·license plates are world-wide standards: every vehicle around the World should have a license plate
·license plates are visible, the existance or lack of a license plate is evident
·license plates are human readable
(OK, not everybody can read Japanese or Saudi number plates... Still, a license plate recognition system has to.)
And what is on the dark side? The answer is basically one single word: accuracy.
Is a license plate recognition system accurate enough for the spacific task? Every recognition system has error, such as
image misinterpretation, misreading, etc. Some applications require very high accuracy, while others can be satisfied with far less.
|
Our products are
ISO 9001:2000 Certified
|
