1

I would like to know how I can process an image of a License plate so I can use this output with an OCR Library like Tesseract. Here you can see an example of the image i'm trying to process http://s13.postimage.org/knf9zvxee/vaflo.jpg

What I am trying to get in this case is just the String "KND 486". So I need to isolate this from everything. You might notice that the string that I am looking for is the biggest one, and as it is a License plate from Argentina, it will always appear 3 letters and then 3 numbers.

As you can see, I've tried to binarize the original picture, but I am not getting the results I expected. Does anyone knows what else should I do?

blasrodri
  • 448
  • 5
  • 16
  • What are the results you expect? Can you elaborate, please? – zkan Jan 17 '13 at 08:34
  • I've edited my question. Please let me know if any other information would be useful. I am know to OpenCV, so I know maybe the question is to broad. – blasrodri Jan 17 '13 at 23:26
  • @blas: Your image is actually pretty clean. You could use blob detection and a simple SVM to classify those letters. – Blender Jan 17 '13 at 23:31
  • Can you tell me how? I am totally new at Open CV – blasrodri Jan 18 '13 at 16:29
  • You may consider a simple algorithm such as shape matching. I hope this http://stackoverflow.com/questions/8785664/finding-shapes-in-an-image-using-opencv helps. – zkan Jan 19 '13 at 12:32

0 Answers0