how to find the threshold value of any image to extract text from credit card image

Question

I am applying the threshold value manually

I am trying to extract exact information of credit card.

I attached the sample images.

how can in extract credit card numbers exactly.

import re

import cStringIO
import urllib

import cv2
import numpy as np
from matplotlib import pyplot as plt
from PIL import Image
from noiselevel import estimate_noise
import pytesseract
oriimage="sh.png"
newimage = cv2.resize(oriimage,(583,327))
newimage = newimage[150:250, 0:600]
  # valid_image = newimage[150:250, 0:600]
cv2.imwrite("moon.tif",newimage)
 # filter=ImageFilter.UnsharpMask(6.0,269,0)
 #  im=Image.open("moon.tif")
 # i=im.filter(filter)
 # i.save("l3.tif")
a=Image.open("moon.tif")
inputt="moon.tif"
img = cv2.imread(inputt,0)

ret1,th1 = cv2.threshold(img,120,254,cv2.THRESH_BINARY)

Here code in CPP, but it not so long, you can easily port it to python: https://stackoverflow.com/questions/22122309/opencv-adaptive-threshold-ocr/22127181#22127181 — Andrey Smorodov, Aug 26 '17 at 09:23
Give us your input image so that we can see what can be done. — I.Newton, Aug 26 '17 at 11:19

score 0 · Answer 1 · answered Aug 27 '17 at 11:56

You can use Adaptive Thresholding to binarise the image. Adaptive thresholding automatically determines the threshold value for given part of an image based on its surroundings. One advantage is that you do not have to hardcode the threshold. Another and more important advantage is that it gives a reasonable binarisation even when there is uneven lighting in the image.

score 0 · Answer 2 · answered Sep 09 '17 at 12:14

0

i improve the accuracy by resize the cropped image.

cv2.resize(newimage,(583,150)

answered Sep 09 '17 at 12:14

how to find the threshold value of any image to extract text from credit card image

2 Answers2