I have a similar project -- to recognize numbers of people's jerseys during a race.
I already have code that detects the area with the number, and tried to use ocropus to do OCR on that part. If I do "ocroscript recognize image.png" on the picture that just contains bib number by itself, I either get nothing or, cryptic errors like "background appears white". The most success I got was by embedding the number I extracted into page with some text, like in picture below http://yaroslavvb.com/upload/run-run/test.png In that case, "bla bla bla" is recognized successfully, and 1016 is recognized as "M6". If I threshold that image and feed in http://yaroslavvb.com/upload/run-run/test2.png, ocropus recognizes "1016" as "@6" OCR is usually improved by restricting the size of the vocabulary. Is there on option to tell Ocropus that output must be digits only? Yaroslav On Jan 14, 5:12 pm, bnaef <[email protected]> wrote: > I am looking for an open-source license plate recognition software to > analyze gray-scale jpg images. > > Has anybody applied OCRopus to a license plate recognition > application? > If yes, would you mind sharing the experience? > If not, could OCRopus serve as a suitable solution for a lpr > application? > > Any other suggestions and/or pointers? > > There are some inquiries on tesseract re this topic, but I couldn't > find anything in this group. (Maybe I missed it). > > With kind regards > Beat Naef --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "ocropus" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/ocropus?hl=en -~----------~----~----~----~------~----~------~--~---
