Textbook-Arrange/main.py

import cv2
import pytesseract
import os
directory = 'images'
 
k = 1
# iterate over files in
# that directory
for filename in os.listdir(directory):
    f = os.path.join(directory, filename)
    # checking if it is a file
    if os.path.isfile(f):
        image = cv2.imread(f, 0)
        thresh = 255 - cv2.threshold(image, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]
        #bottom 2 locations
        ROI1 = thresh[1383:1447,77:154] 
        ROI2 = thresh[1383:1447,950:1027]
        ROI = cv2.hconcat([ROI1, ROI2])
        data = pytesseract.image_to_string(ROI, lang='eng',config='--psm 7 -c tessedit_char_whitelist=0123456789i' )
        print(data)
        if os.path.isfile(f'images/{data}.jpg') == False:
            os.rename(f'{f}', f'images/{data}.jpg')
        else:
            os.rename(f'{f}', f'images/{data}.{k}.jpg')
            k = k + 1
        cv2.destroyAllWindows()