Looking to purchase source code for license plate recognition

Thread Starter

lukied123

Joined Aug 30, 2011
6
Hey everyone,

I am looking to purchase source code that can identify and recognize a license plate within a digital picture. It would have to convert a given license plate within a picture to alphanumeric ASCII and be able to successfully identify all characters on the plate. Anyone interested?
 

someonesdad

Joined Jul 7, 2009
1,583
Why don't you specify the problem in more detail? Questions that pop to mind are:

What is the application?
What decisions will be made with the data?
What is the desired error rate in fraction of characters in error per character?
What's the range of input images? Can you give an example of a typical picture and a few examples of worst-case images?
Are the images monochromatic or will they have a variety of colors?
What's the budget?
What's the timeline?
How fast does the conversion have to be done?
What platform(s) does the application(s) have to run on?
Describe the computing: is it a batch-type environment? Is there a user interacting with a UI or do the images e.g. come over a socket or some other method?
 

maxpower097

Joined Feb 20, 2009
816
First off try to find some open source OCR. Once you have that then port it to your embedded platform. Heres a lil head start This one seems to be the best free one with code available. They have it running on Linux and Win, so an android or small linux system would be ideal.
http://code.google.com/p/tesseract-ocr/
Heres a link to the actual source code
http://code.google.com/p/tesseract-ocr/downloads/detail?name=tesseract-3.00.tar.gz&can=2&q=

Ideally you would set up your camera and figure out a trigger to take the shot. If its at a stop light may be red lights like brake lights could trigger the snapshot. Then take the snapshot of the plate and run it through the OCR. Then it should spit out all the text in the shot.
 
Last edited:
Top