Hi ,
Recently i began using your Cloud OCR SDK demo example to test the OCR capabilities.
It worked really well on text documents etc but would return blank when trying to return the text out of an image i.e a vehicle number plate. Is there a particular tool i should be using to OCR a cars licence plate? here is an example of what i am trying to do: http://www.burningwell.org/gallery2/d/7583-3/dscf0204.jpg. Basically i would want to return the licence plate of this vehicle as a string .
I need to be able to do this prefereably using Javascript also.
Thanks in advance.
コメント
4件のコメント
To recognize license plates it is necessary to use field-level recognition (in this case text coordinates are specified directly).
Hi,
Thanks for your reply! Is there a working example of this method available to test in JavaScript?
Or can i simply modify the current process image example to get processTextField method working?
Appreciate the Help.
Used this Method suggested in C# and found the accuracy on car license plates to not be very accurate. also having to specify coords/textfield isn't quite what i need as people using the tool will be on a mobile phone and as a result the images taken will be from all sorts of heights/angles. Is their something/a method i have missed? Perhaps something that will search the image and find the text without having to specify a region? Thanks again.
Unfortunately, we have no automatic analysis for license plates.
To achieve the best recognition quality, we recommend to specify the following parameters of the processTextField method:
textType ("normal", while the default value is "normal,handprinted");
(optional) letterset;
(optional) regExp.
The description of all the parameters is here: http://ocrsdk.com/documentation/apireference/processTextField/
サインインしてコメントを残してください。