Problema with convert image to string

[karlo123]

Hi everyone,

I am working on decoding a text of an image, for this I take a complicated case, captcha with lines ... I imagine that if it works well it will do it in fuzzy texts or in scratched documents ...

The problem is that when I decode, the result is very different from what I'm trying to read, why is this?

    ##############################
    import pytesseract
    from PIL import Image
    
    import cv2
    import numpy as np
    
    img = Image.open('C:/XYZ/AAAA/captcha2_Ejemplo.bmp').convert('RGB') 
    
    ##############################
    cv_img = np.asarray(img)[:,:,::].copy()
    
    gray = cv2.cvtColor(cv_img, cv2.COLOR_BGR2GRAY)
    
    edges = cv2.Canny(gray, 60, 200, apertureSize = 3)
    
    lines = cv2.HoughLinesP(edges, 1, np.pi / 180, 1, None, 0, 0)
    
    for x1, y1, x2, y2 in lines[0]:
        cv2.line(cv_img, (x1, y1), (x2, y2), (255,255,255 ), 2)
        
    processed = cv2.cvtColor(cv_img, cv2.COLOR_BGR2GRAY)

    blur = cv2.GaussianBlur(processed, (3, 3), 0)

    threshold = cv2.threshold(blur, 128, 255, cv2.THRESH_BINARY)[1]

    ##############################

    pil_img = Image.fromarray(threshold)
    
    pytesseract.pytesseract.tesseract_cmd = 'C:/Program Files (x86)/Tesseract-OCR/tesseract'
    result = pytesseract.image_to_string(threshold)
    print(result)  
    ##############################    

Image for example:
https://drive.google.com/open?id=1oU4Qom...Uh-YXaLE4l

The result is "Ky é-Tel 3"... the correct result is "32YSQE"...

I will greatly appreciate your help


(Main source: https://gist.github.com/chroman/8485394)

Thanks
Karlo

[karlo123]

Any help?...please