[FIXED] Improving the tesseract detection

Issue

I tried to extract text from this image using Tesseract.

The code that I tried:

img = Image.open('downloadedpng.jpeg').convert('L')
ret,img = cv2.threshold(np.array(img), 125, 255, cv2.THRESH_BINARY)
img = Image.fromarray(img.astype(np.uint8))
print(pytesseract.image_to_string(img))

The output that I got: re vie
I've tried erosion and dilation with the below code:

img_erosion = cv2.erode(img, kernel, iterations=1)
img_dilation = cv2.dilate(img, kernel, iterations=1)

but I got errors. Any idea how to properly convert it to string ?

Solution

You should pass a black on white text for best results:

import cv2
from PIL import Image

img = cv2.imread('1.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
img = cv2.bitwise_not(img) # <- invert
ret, thresh = cv2.threshold(img, 100, 255, cv2.THRESH_BINARY)
im = Image.fromarray(thresh.astype("uint8"))
print(pytesseract.image_to_string(im))

Answered By - K41F4r