常见的验证码分为三类:
本文重点讲解前两种的自动化处理方案。
ddddocr 是目前最好用的开源验证码识别库,离线运行、开箱即用,对常规4-6位字母数字验证码准确率可达99%以上。
pip install ddddocr requestsimport ddddocr
ocr = ddddocr.DdddOcr(show_ad=False)
with open("captcha.png", "rb") as f:
result = ocr.classification(f.read())
print("识别结果:", result)import requests
import ddddocr
ocr = ddddocr.DdddOcr(show_ad=False, det=True) # det=True 开启增强模式
code_url = "https://wlt.hbguining.com/captcha.png"
resp = requests.get(code_url)
if resp.status_code == 200:
result = ocr.classification(resp.content)
print("识别结果:", result)import time
import ddddocr
from selenium import webdriver
from selenium.webdriver.common.by import By
driver = webdriver.Chrome()
driver.get("https://wlt.hbguining.com/login")
ocr = ddddocr.DdddOcr(show_ad=False, det=True)
# 获取验证码图片并识别
captcha_img = driver.find_element(By.ID, "captcha-img")
captcha_img.screenshot("captcha.png")
with open("captcha.png", "rb") as f:
code = ocr.classification(f.read())
print("识别验证码:", code)
# 填写表单
driver.find_element(By.NAME, "username").send_keys("admin")
driver.find_element(By.NAME, "password").send_keys("123456")
driver.find_element(By.NAME, "captcha").send_keys(code)
driver.find_element(By.ID, "login-btn").click()
time.sleep(3)
driver.quit()滑块验证码的核心是两步:定位缺口位置 + 模拟真人轨迹。
pip install captcha-recognizer selenium numpy使用 captcha-recognizer 库,两行代码即可定位缺口坐标:
from captcha_recognizer.slider import Slider
slider = Slider()
box, confidence = slider.identify(source="slide_bg.png")
print(f"缺口坐标: {box}, 置信度: {confidence}")
# box = [x1, y1, x2, y2],取 x1 即为缺口水平位置
gap_x = box[0]风控系统会检测鼠标轨迹,匀速直线移动必被拦截。需要模拟加速-减速+随机抖动的真人轨迹:
import random
import time
from selenium.webdriver import ActionChains
def generate_track(distance):
"""生成真人滑动轨迹"""
track = []
current = 0
mid = distance * 0.6 # 前60%加速,后40%减速
speed = 0
while current < distance:
if current < mid:
speed += random.uniform(1.5, 2.5)
else:
speed -= random.uniform(1.0, 2.0)
speed = max(0.5, speed)
step = min(int(speed), distance - current)
current += step
y_offset = random.randint(-2, 2) # 垂直微抖
track.append((step, y_offset))
# 末端轻微回弹,模拟真人对齐误差
track.append((-random.randint(2, 5), random.randint(-1, 1)))
return track
def slide(driver, slider_element, distance):
"""执行滑块拖动"""
actions = ActionChains(driver)
actions.click_and_hold(slider_element).perform()
time.sleep(0.3)
track = generate_track(distance)
for dx, dy in track:
actions.move_by_offset(dx, dy).perform()
time.sleep(random.uniform(0.01, 0.03))
time.sleep(0.5)
actions.release().perform()import time
import requests
from selenium import webdriver
from selenium.webdriver.common.by import By
from captcha_recognizer.slider import Slider
driver = webdriver.Chrome()
driver.get("https://wlt.hbguining.com/login")
# 1. 获取滑块背景图
bg_url = driver.find_element(By.CLASS_NAME, "slide-bg").get_attribute("src")
bg_data = requests.get(bg_url).content
with open("slide_bg.png", "wb") as f:
f.write(bg_data)
# 2. 识别缺口位置
slider = Slider()
box, _ = slider.identify(source="slide_bg.png")
gap_x = box[0]
# 注意:某些验证码存在初始偏移量,需减去偏移
# offset, _ = slider.identify_offset(source="slide_bg.png")
# gap_x = gap_x - offset
# 3. 执行滑动
slider_btn = driver.find_element(By.CLASS_NAME, "slide-btn")
slide(driver, slider_btn, gap_x)
time.sleep(2)
driver.quit()场景 | 推荐方案 | 特点 |
|---|---|---|
简单图形验证码 | ddddocr | 离线、免费、高准确率 |
滑块验证码 | captcha-recognizer + 轨迹模拟 | 需处理坐标偏移和轨迹 |
reCAPTCHA / 极验 | CapSolver / 2Captcha | 付费API,省心但花钱 |
以上技术仅限用于个人学习、自有系统自动化测试、授权合规的数据采集。禁止用于恶意爬取第三方平台、批量破解账号等违规行为,可能涉及法律风险。
原创声明:本文系作者授权腾讯云开发者社区发表,未经许可,不得转载。
如有侵权,请联系 cloudcommunity@tencent.com 删除。