search

openalpr / train-ocr

Input files and scripts necessary to train the license plate OCR
GNU Affero General Public License v3.0
234 stars 109 forks source link

Help to train OCR #8

Closed Barathlakshmanan closed 9 years ago

Barathlakshmanan commented 9 years ago

I am training openALPR for Chinese characters. I have problem while training my images, the code accepts only A-Z and 0-9 as inputs, can you guide me how to give Chinese characters as input? I have scim installed to switch between languages but i am not able to do it with this trainer. Any help is appreciated. Thanks in advance.

matthill commented 9 years ago

Yes, you're right. The training program only accepts ASCII keys as inputs. Tesseract and OpenALPR both support Unicode. This trainer will need to get updated as well.

matthill commented 9 years ago

I've made enhancements to the training workflow when adding support for South Korean license plate OCR. China should be similar.

I used a TTF font sheet to do the training. I added some documentation describing this process.

xiebianhuzi commented 7 years ago

@Barathlakshmanan have you solve the problems, please tell me how to train openalpr for Chinese characters,thank you,or any documentation describing the process?

xiebianhuzi commented 7 years ago

@matthill have you maked it?can you tell me the detailed process?thank you.I want to train openALPR for Chinese characters.

xiebianhuzi commented 7 years ago

@matthill @Barathlakshmanan I need some help,please,my mailbox is xiebianhuzi@gmail.com,thank you.