Help with detecting a code on an image

[nabil-ak]

Hello,

I want to detect a code (combination of characters) on an image in python.

I already tried EasyOCR and Tesseract Open Source OCR Engine but noone could detect the characters.

I also tried to preprocess the picture by inverting the white font to a black font and painting the background white to make it for the engine easier to detect the characters. 

What am i doing wrong?

[Umanda Dikwatta]

Did you try all the preprocessing steps? 

Rotation / Deskewing ? I think Tesseract finds it difficult to identify skewed images.

--
You received this message because you are subscribed to the Google Groups "tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-oc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/2e764275-4992-495f-b0ac-ffe668254231n%40googlegroups.com.

[nabil-ak]

I also tried Rotation but its still not working.

These are the preprocessing steps that i used:

import cv2

import pytesseract

import numpy as np

from scipy import ndimage

img = cv2.imread('voucher.png')

img = cv2.bitwise_not(img)

img = cv2.resize(img, None, fx=1.2, fy=1.2, interpolation=cv2.INTER_CUBIC)

img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

kernel = np.ones((1, 1), np.uint8)

img = cv2.dilate(img, kernel, iterations=1)

img = cv2.erode(img, kernel, iterations=1)

img = ndimage.rotate(img, -20)

cv2.imwrite("changed.png",img)

pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'

print(pytesseract.image_to_string(img))

[nabil-ak]

https://imagetotext.info/

If i use a website like this one it can extract the text perfectly and they also use tesseract.

There has to be some preproccesing/setting that makes tesseract detect the text perfectly.

[kwmz...@gmail.com]

So I think you could also try using some morphological transformation; I feel like dilation could help a lot too 

[Nikhil Fande]

Hi,

Combination of image deskewing and image inversion should work.

Use PSM 11 on processed image, it will detect most of the characters.

Thanks,

Nikhil

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/e49c5a3e-b810-4714-92e8-1b78d03f5e40n%40googlegroups.com.

[nabil-ak]

Thanks for all the help.

I tried closing as a morphological transformation and also used a threshold to get a black and white picture.

I got "BYC3 .- A4UFYY.QLBGY -+ " as a answer this is almost right.

Is there anything else that i can do to improve the detection?

i attached the python code that i used and the picture after the preprocessing.

import cv2

import pytesseract

import numpy as np

from scipy import ndimage

img = cv2.imread('voucher.png')

img = cv2.bitwise_not(img)

img = cv2.resize(img, None, fx=2, fy=2, interpolation=cv2.INTER_CUBIC)

img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

_,img = cv2.threshold(img,145,255,cv2.THRESH_BINARY)

kernel = np.ones((2, 2), np.uint8)

img = cv2.dilate(img, kernel, iterations=1)

img = cv2.erode(img, kernel, iterations=1)

img = ndimage.rotate(img, -20)

cv2.imwrite("changed.png",img)

pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'

print(pytesseract.image_to_string(img))