How do i get accurate character segmentation In ANPR?

asked 2019-12-18 04:00:57 -0500

updated 2019-12-18 04:46:05 -0500

I was trying to do a License Plate recognition project.I was able to successfully extract license plates from the images.

However, the character segmentation step is giving unsuccessful results in many cases.

I have tried a variety of ways including various methods using OPEN CV and OCR. But most of them seem to miss out a few characters or give totally unusable results.

Please help me with a good approach or a reference for a robust solution to the problem.

Sample Image -

Sample Image

    img_gray=cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
_,img_binary=cv2.threshold(img_gray,200,255,cv2.THRESH_BINARY + cv2.THRESH_OTSU)

img_erode = cv2.erode(img_binary, (3,3))
img_dilate = cv2.dilate(img_erode, (3,3))

LP_WIDTH = img_dilate.shape[0]
LP_HEIGHT = img_dilate.shape[1]

dimensions = [LP_WIDTH/6, LP_WIDTH/2, LP_HEIGHT/10, 2*LP_HEIGHT/3]

char_list=find_contours(dimensions, img_dilate)

return char_list

Thanks in advance,

Michael

edit retag flag offensive close merge delete

Comments

I have tried a variety of ways

show please, else it's hot air

berak gravatar imageberak ( 2019-12-18 04:14:43 -0500 )edit