search

tesseract-ocr / tesseract

Tesseract Open Source OCR Engine (main repository)
https://tesseract-ocr.github.io/
Apache License 2.0
61.96k stars 9.48k forks source link

Plate Detection Failing #1214

Closed mmderakhshani closed 6 years ago

mmderakhshani commented 6 years ago

Hi, I have installed tesseract version 4 (LSTM-based OCR) and Leptonika-1.74.4 on my ubuntu 16.04. I have decided to detect the character on a license plate.

plak

I have run bellow command:

tesseract plak.jpeg output --oem 1 -l fas

But the output is not appropriate.

٤ ت
ات

۵٣ 6۱

Just some characters without any logic. Could you please help me how I can get a better result for this license. Another thing which I would like to mention is that the license is related to (Farsi/Persian) language.

Best, Mohammad

roozgar commented 6 years ago

​for this kind of detection you can train your own fonts because license plate in iran have some limited characters ​...

mmderakhshani commented 6 years ago

@roozgar, Thanks for your great response. Could you please give some more information?

roozgar commented 6 years ago

its better to ask this kind of question in google groups because here a developer repository how ever you need image processing to split the pelak in image and detect letter in each image...