as you can see in the title, Im looking for a solution that can make a real-time license plate recognation.
i have been able to make it work using my own code ( yolo and darknet ) to detect the type of the vehicle ( car, bus, moto ) and to detect the plate in theses vehicule.
i have been also to do my own image segmentation for the chars and performe ocr using sklearn ( python ).
however my success rate is 75% ( success rate of correct reading for the numbers and letters -OCR , and not the rate of detection ) due to dotted letter like some arabic letters and English one ( i , j ) my segmetation just ignore the dotes.
but I'm not stusfied with the result.
so here what i need in details:
- a none-GPU based solution ( my realtime will be just 10 frames per seconds , so no need to do a gpu implementation )
- the detection of the car and the plate will be in an arduino so better to be with python or C ( i can share my detection if you think that gonna help you, but only if you will do the rest)
- the image of the plate will be send to a server where will be read by the system ( so the segmentation the recognition ... ) ,and i insist no GPU need.
-the exchange can be done via Webservice or Websocket, at this point i don't care about the UI or the mechanism , i can edit your code by my own so if you think that adding the connection and the distributed system gonna just lose your time and my budget so don't.
-i need a success rate over 85% at least ,90% will be better.
- some arabic Alphabet will be included , so English Alphabet, numbers [0-9] , arabic Alphabet , and spliter like - / | \.
im not un a hurry so take your time with that.
i can send you some samples of plates and also some video feed to see if your existing code can be improved to reach my expectations.
Bu iş için 3 freelancer ortalamada €197 teklif veriyor
Hello, sir I am a professional image processing developer I have my ALPR engine in c++. The speed of my engine is very fast. So you can use it in any platform and CPU We can discuss more in chat. Best regards