Automating Arithmetic Captcha Solving with Python, Requests, pytesseract, and Selenium

The article describes a method to bypass arithmetic captchas used by a telecom operator by programmatically downloading captcha images, recognizing the expression, calculating the result, and automatically clicking the matching image.

Step 1 – Download captcha images : Using the requests library, multiple captcha pictures are fetched and saved locally. Example code:

import requests

def download_captcha_images(base_url, num_images, save_path):
    for i in range(1, num_images+1):
        captcha_url = f"{base_url}/{i}.png"
        file_path = f"{save_path}/captcha{i}.png"
        try:
            response = requests.get(captcha_url, stream=True)
            if response.status_code == 200:
                with open(file_path, 'wb') as f:
                    for chunk in response.iter_content(1024):
                        f.write(chunk)
                print(f"验证码图片 {i} 保存成功！")
            else:
                print(f"无法获取验证码图片 {i}，状态码：{response.status_code}")
        except requests.exceptions.RequestException as e:
            print(f"请求验证码图片 {i} 时发生异常：{e}")

# 示例：获取多张验证码图片并保存到本地
base_url = "https://example.com/captcha"  # 替换成实际的验证码图片URL基础部分
num_images = 5  # 假设需要获取5张验证码图片
save_path = "captcha_images"  # 保存路径，根据实际情况设置

download_captcha_images(base_url, num_images, save_path)

Step 2 – Recognize and compute the arithmetic expression : The pytesseract OCR library extracts the text from a captcha image, the numbers are parsed, and the sum is calculated. Example code:

from PIL import Image
import pytesseract

def recognize_captcha(image_path):
    try:
        text = pytesseract.image_to_string(Image.open(image_path))
        # 假设验证码格式为 "x + y = ?"
        x, y = map(int, text.split('+'))
        result = x + y
        return result
    except Exception as e:
        print("发生异常：", e)
        return None

# 示例使用
captcha_image = "captcha.png"  # 替换成实际的验证码图片路径
result = recognize_captcha(captcha_image)
if result is not None:
    print(f"验证码算式的结果为：{result}")
else:
    print("无法识别验证码，请检查验证码图片是否有效。")

Step 3 – Automate verification and clicking with Selenium : Selenium drives a browser to load the target page, captures each captcha image, runs the OCR routine, compares the computed result with a condition, and performs a click when the condition is met. Example code:

from selenium import webdriver
from PIL import Image
import pytesseract

def recognize_captcha(image_path):
    try:
        text = pytesseract.image_to_string(Image.open(image_path))
        x, y = map(int, text.split('+'))
        return x + y
    except Exception as e:
        print("发生异常：", e)
        return None

def main():
    url = "https://example.com"  # 替换成实际的网站URL
    driver = webdriver.Chrome()  # 需要对应版本的 ChromeDriver
    driver.get(url)
    for i in range(1, 6):  # 假设有 5 张验证码图片
        captcha_element = driver.find_element_by_xpath('//*[@id="captcha_image"]')
        captcha_image = captcha_element.screenshot_as_png
        captcha_image_path = f"captcha{i}.png"
        with open(captcha_image_path, "wb") as f:
            f.write(captcha_image)
        captcha_result = recognize_captcha(captcha_image_path)
        if captcha_result is not None:
            print(f"验证码{i}算式的结果为：{captcha_result}")
            if captcha_result == 10:
                submit_button = driver.find_element_by_xpath('//*[@id="submit_button"]')
                submit_button.click()
                print(f"验证码{i}满足条件，执行点击操作。")
        else:
            print(f"验证码{i}无法识别，请检查验证码图片是否有效。")
    driver.quit()

if __name__ == "__main__":
    main()

Additional notes remind readers to handle possible anti‑scraping measures, install required drivers, respect website terms of service, and consider more advanced image‑processing or deep‑learning techniques for complex captchas.