Python3.7爬虫：实时api(百度ai)检测验证码模拟登录(Selenium)页面

temp_url = 'https://aip.baidubce.com/rest/2.0/ocr/v1/accurate_basic?access_token=' + token
temp_headers = {'Content-Type': 'application/x-www-form-urlencoded'}
temp_file = open('./test.png', 'rb')
temp_image = temp_file.read()
temp_file.close()
temp_data = {'image': base64.b64encode(temp_image)}
temp_data = urllib.parse.urlencode(temp_data)
temp_res = requests.post(url=temp_url, data=temp_data, headers=temp_headers)
res = json.loads(str(temp_res.text))
code = res['words_result'][0]['words']

#删除元素
js = "var elem = document.getElementsByClassName('dnf');elem[0].parentNode.removeChild(elem[0]);"
driver.execute_script(js)
js = "var elem = document.getElementsByClassName('qqshow');elem[0].parentNode.removeChild(elem[0]);"
driver.execute_script(js)
js = "var elem = document.getElementsByClassName('QR');elem[0].parentNode.removeChild(elem[0]);"
driver.execute_script(js)

这里还有个技巧，可以提高百度api的识别准确度，就是利用opencv对验证码图片进行降噪去色的操作。

首先安装opencv

pip3 install opencv-python

然后对原图进行去色处理

import cv2

img = cv2.imread('./code.png', cv2.IMREAD_GRAYSCALE)
print("cv2.imread(imgfile, cv2.IMREAD_GRAYSCALE)结果如下：")
print('大小：{}'.format(img.shape))
print("类型：%s"%type(img))
print(img)

cv2.imwrite('code1.png', img)

去色降噪后的效果是这样的，可以提高识别成功率

当然了，灰度处理可能有时候还不能达到标准，也可以进一步进行二值化处理：

import cv2
import matplotlib.pyplot as plt
img = cv2.imread("/Users/liuyue/Downloads/code.png")
img2 = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
img2 = cv2.inRange(img2, lowerb=160, upperb=255)
plt.subplot(121), plt.imshow(img)  # 原始图片
plt.subplot(122), plt.imshow(img2)  # 降噪图片
plt.show()

selenium驱动下载地址：http://npm.taobao.org/mirrors/chromedriver/

剩下的就很简单了，直接利用selenium输入账号密码以及验证码就可以自动登录，非常轻松，完整代码如下：

import requests
import json
import base64
import urllib
from selenium import webdriver
import time
#selenium 截图
driver = webdriver.Chrome()
driver.get('https://www.dianxiaomi.com/index.htm')
time.sleep(1)
#删除元素
js = "var elem = document.getElementsByClassName('dnf');elem[0].parentNode.removeChild(elem[0]);"
driver.execute_script(js)
js = "var elem = document.getElementsByClassName('qqshow');elem[0].parentNode.removeChild(elem[0]);"
driver.execute_script(js)
js = "var elem = document.getElementsByClassName('QR');elem[0].parentNode.removeChild(elem[0]);"
driver.execute_script(js)
nodes = driver.find_element_by_id('loginImgVcode')
nodes.screenshot('test.png')
#用户名 exampleInputName exampleInputPassword
driver.find_element_by_id('exampleInputName').send_keys('18231327142')
driver.find_element_by_id('exampleInputPassword').send_keys('密码')
#请求百度api接口
res = requests.get("https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=你的key&client_secret=你的秘钥")
res = json.loads(str(res.text))
token = res['access_token']
temp_url = 'https://aip.baidubce.com/rest/2.0/ocr/v1/accurate_basic?access_token=' + token
temp_headers = {'Content-Type': 'application/x-www-form-urlencoded'}
temp_file = open('./test.png', 'rb')
temp_image = temp_file.read()
temp_file.close()
temp_data = {'image': base64.b64encode(temp_image)}
temp_data = urllib.parse.urlencode(temp_data)
temp_res = requests.post(url=temp_url, data=temp_data, headers=temp_headers)
res = json.loads(str(temp_res.text))
print(res)
code = res['words_result'][0]['words']
driver.find_element_by_id('loginVerifyCode').send_keys(code.strip().replace(' ',''))
#登录按钮
driver.find_element_by_id('loginBtn').click()