With a younger brother OCR, say no to various types of verification codes!

writing |  So and so white rice

source ：Python technology 「ID: pythonall」

When using a crawler to log in to a website , You often encounter verification code after entering user name and password , A simpler code with letters , A little more complicated is the slider verification code , Click the article and click the picture verification code . These are all chronic problems in reptiles , Today, we will introduce a universal verification code SDK Say goodbye to them completely , Its name is ddddocr Take your brother OCR Universal verification code identification SDK Free open source .

install

Will automatically install the latest software that conforms to your computer environment ddddocr.Python Environment needs to be less than or equal to 3.9.

pip install ddddocr

Use

Take your brother OCR Three verification codes can be identified , Xiaobian will experiment with these three methods .

Slider verification code

Slider verification code here is the slider verification of Douban . The lower slider is a separate transparent background hycdn.png.

hycdn.png

The background image is a pit with a small slider background.jpg.

background

import ddddocr

det = ddddocr.DdddOcr(det=False, ocr=False)

with open('hycdn.png', 'rb') as f:
        target_bytes = f.read()
    
with open('background.jpg', 'rb') as f:
    background_bytes = f.read()

res = det.slide_match(target_bytes, background_bytes, simple_target=True)

print(res)

Recognition result

{'target_y': 0, 'target': [486, 126, 622, 262]}

target The first two values of the attribute coincide with the Douban validation slider url Submitted in ans almost .

Click the class verification code

Click the type verification code to use Netease login

eb.jpg

det = ddddocr.DdddOcr(det=True)

    with open("eb.jpg", 'rb') as f:
        image = f.read()

    poses = det.detection(image)

    im = cv2.imread("eb.jpg")

    for box in poses:
        x1, y1, x2, y2 = box
        im = cv2.rectangle(im, (x1, y1), (x2, y2), color=(0, 0, 255), thickness=2)

    cv2.imwrite("result.jpg", im)

Recognition result

Alphanumeric verification code

The picture of alphanumeric verification code comes from google Search for

ocr = ddddocr.DdddOcr(old=True)

with open("z1.jpg", 'rb') as f:
    image = f.read()

res = ocr.classification(image)
print(res)

Recognition result

3n3d
8342

summary

ddddocr Make the verification code so simple and easy to use ,, It won't work opencv, pytorch, tensorflow Our little partner can also quickly crack the login verification code of the website . If you have other good friends ocr Recognition can also be shared in messages .

PS： Reply within company number 「Python」 You can enter Python Novice learning communication group , Together 100 Day plan ！

Old rules , Brothers, remember , Right lower corner “ Looking at ” click , If you think the content of the article is good , Remember to share the circle of friends and let more people know ！

【 Code Access method 】

Identify the QR code at the end of the text , reply ： So and so white rice