百度百家号旋转验证码识别代码分享

最近研究了一下图像识别，一直找到很好的应用场景，今天我就发现可以用百度的旋转验证码来做一个实验。没想到效果还挺好，下面就是实际的识别效果。

1、效果演示

2、如何识别

2.1准备数据集

首先需要使用爬虫，对验证码图片进行采集，尽量每一种类型都要采集到。

2.2图像矫正

接下来对采集的数据进行人工校正

2.3数据清洗

（1）对数据进行进行旋转，达到增加数据量的目的。

（2）对数据进行灰度化处理，将三维图片降为二维。

（3）对图片大小进行resize，可以提高训练速度。

# 图片转换部分，得到x
picture = Picture(path=img_path)
# 图像灰度化处理
temp_img = picture.gray()
# 图像resize
temp_img = temp_img.resize((50, 50), Image.LANCZOS)
# 获取y
word = img_path.split('\\')[-1].split('-')[0]
 
# 结果包装成列表，保证x，y是一个整体，不被打乱
res = [np.array(temp_img),np.array(word)]
# 将结构给全局变量
result_list.append(res)
# 记录完成数量
complete_list.append(img_path)

2.4划分训练集与测试集

一般训练集占数据量的80%，测试集占总数据量的20%，当然也可以根据自己的情况调整比例。

2.5训练模型

这里可以使用CNN神经网络模型进行训练，效果非常不错。

2.6实战测试

下面直接上代码。其中的滑动系数可能需要自行调整，这个变动不会太频繁，可能几个月某度变一次。

__author__ = "dengxinyan"
 
import os
import sys
import time
import base64
import random
import requests
from PIL import Image
from io import BytesIO
sys.path.append(os.path.abspath(os.path.dirname(os.path.abspath(os.path.dirname(__file__)))))
from selenium import webdriver
from selenium.webdriver.support.wait import WebDriverWait
from selenium.webdriver import ActionChains
 
#PIL图片保存为base64编码
def PIL_base64(img, coding='utf-8'):
    img_format = img.format
    if img_format == None:
        img_format = 'JPEG'
 
    format_str = 'JPEG'
    if 'png' == img_format.lower():
        format_str = 'PNG'
    if 'gif' == img_format.lower():
        format_str = 'gif'
 
    if img.mode == "P":
        img = img.convert('RGB')
    if img.mode == "RGBA":
        format_str = 'PNG'
        img_format = 'PNG'
 
    output_buffer = BytesIO()
    # img.save(output_buffer, format=format_str)
    img.save(output_buffer, quality=100, format=format_str)
    byte_data = output_buffer.getvalue()
    base64_str = 'data:image/' + img_format.lower() + ';base64,' + base64.b64encode(byte_data).decode(coding)
 
    return base64_str
 
# 根据链接下载旋转图片
def get_img(url):
    header = {
        "Host": "passport.baidu.com",
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:72.0) Gecko/20100101 Firefox/72.0",
        "Accept": "image/webp,*/*",
        "Accept-Language": "zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2",
        "Accept-Encoding": "gzip, deflate, br",
        "Connection": "keep-alive",
        "Referer": "https://wappass.baidu.com/static/captcha/tuxing.html?&ak=c27bbc89afca0463650ac9bde68ebe06&backurl=https%3A%2F%2Fwww.baidu.com%2Fs%3Fcl%3D3%26tn%3Dbaidutop10%26fr%3Dtop1000%26wd%3D%25E6%25B6%2588%25E9%2598%25B2%25E6%2588%2598%25E5%25A3%25AB%25E8%25BF%259E%25E5%25A4%259C%25E7%25AD%2591%25E5%259D%259D%25E5%25BA%2594%25E5%25AF%25B9%25E6%25B4%25AA%25E5%25B3%25B0%25E8%25BF%2587%25E5%25A2%2583%26rsv_idx%3D2%26rsv_dl%3Dfyb_n_homepage%26hisfilter%3D1&logid=8309940529500911554&signature=4bce59041938b160b7c24423bde0b518&timestamp=1624535702",
        "Cookie": "BAIDUID=A0621DC238F4D936B38F699B70A7E41F:SL=0:NR=10:FG=1; BIDUPSID=A0621DC238F4D9360CD42C9C31352635; PSTM=1667351865; HOSUPPORT=1; UBI=fi_PncwhpxZ%7ETaKAanh2ue0vFk6vHMY02DgvigILJIFul8Z1nzMr9do3SYLtjAUqHSpUz7LvOKV27cIr18-YJryP0Q8j92oo93%7E6hGa0CLdraAlaHUZG-0PW9QrpZkW7MTyUn-yrAq7OmSRBIJ7%7E8gM9pv-; USERNAMETYPE=2; SAVEUSERID=3cd458184c56c2fe28174e594101f074d63463446d; HISTORY=0ece87e30ec8ecccd52ff3d5c42f98002a893bfb73ff358893; BDUSS_BFESS=NOcWd6YWJRbmFVUVBBaWVkaHJNSm5tRUpUaUVMaTNHOHcwZVVaVDdsYXlLZmxrSVFBQUFBJCQAAAAAAAAAAAEAAAC13Mct0KHQwl9keHkAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAALKc0WSynNFkRD; H_WISE_SIDS=219946_216846_213346_219942_213039_230178_204909_230288_110085_236307_243888_244730_245412_243706_232281_249910_247148_250889_249892_252577_234296_253427_253705_240590_254471_179345_254689_254884_254864_253213_255713_254765_255939_255959_255982_107317_256062_256093_256083_255803_253993_256257_255661_256025_256223_256439_256446_254831_253151_256252_256196_256726_256739_251973_256230_256611_256996_257068_257079_257047_254075_257110_257208_251196_254144_257290_251068_256095_257287_254317_251059_251133_254299_257454_257302_255317_255907_255324_257481_244258_257582_257542_257503_255177_257745_257786_257937_257167_257904_197096_257586_257402_255231_257790_258193_258248_258165_8000084_8000115_8000114_8000126_8000140_8000149_8000166_8000172_8000178_8000181_8000185_8000204; ZFY=SxMcCdU3pSsmienZSgA2BTmHLR9S6caVmiP5Ic:Awuz0:C; BAIDUID_BFESS=A0621DC238F4D936B38F699B70A7E41F:SL=0:NR=10:FG=1; Hm_lvt_90056b3f84f90da57dc0f40150f005d5=1690961642,1692328306; STOKEN=01dbff3d6ff696219b39c9fb730c31c34e032c0eebff4fe535d2f1dde0c7b45b; BDUSS=NOcWd6YWJRbmFVUVBBaWVkaHJNSm5tRUpUaUVMaTNHOHcwZVVaVDdsYXlLZmxrSVFBQUFBJCQAAAAAAAAAAAEAAAC13Mct0KHQwl9keHkAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAALKc0WSynNFkRD; PTOKEN=92e828db8120372a7baa2557ea4ec476; MAWEBCUID=web_VYfxPuQDaKjEzVgXMFgoHouACkpXyjcDpcWwhATKqELuuwEtNy; __bid_n=18a4ab547aa11525d249ea",
    }
    response = requests.get(url=url,headers=header)
 
    if response.status_code == 200:
        img = Image.open(BytesIO(response.content))
 
    # 将图片转换成base64字符串并返回
    return PIL_base64(img)
 
# 验证码识别接口
def shibie(img_base64):
    url = "http://www.detayun.cn/openapi/verify_code_identify/"
    data = {
        # 用户的key
        "key":"JxSfP4E1yfcmJZB6ynOB",
        # 验证码类型
        "verify_idf_id":"16",
        # 样例图片
        "img_base64":img_base64,
        "img_byte": None,
        # 中文点选，空间语义类型验证码的文本描述（这里缺省为空字符串）
        "words":""
    }
    header = {"Content-Type": "application/json"}
 
    # 发送请求调用接口
    response = requests.post(url=url, json=data, headers=header)
    print(response.text)
    return int(str(response.json()['data']['res_str']).replace('顺时针旋转','').replace('度',''))
 
 
if __name__ == '__main__':
    # 加载防检测js
    with open('.\webdriver\stealth.min.js') as f:
        js = f.read()
 
    options = webdriver.ChromeOptions()
    driver = webdriver.Chrome(executable_path='.\webdriver\chromedriver.exe', options=options)
 
    driver.execute_cdp_cmd("Page.addScriptToEvaluateOnNewDocument", {
        "source": js
    })
    # 访问百度首页
    driver.get('https://wappass.baidu.com/static/captcha/tuxing.html?&ak=c27bbc89afca0463650ac9bde68ebe06&backurl=https%3A%2F%2Fwww.baidu.com%2Fs%3Fcl%3D3%26tn%3Dbaidutop10%26fr%3Dtop1000%26wd%3D%25E6%25B6%2588%25E9%2598%25B2%25E6%2588%2598%25E5%25A3%25AB%25E8%25BF%259E%25E5%25A4%259C%25E7%25AD%2591%25E5%259D%259D%25E5%25BA%2594%25E5%25AF%25B9%25E6%25B4%25AA%25E5%25B3%25B0%25E8%25BF%2587%25E5%25A2%2583%26rsv_idx%3D2%26rsv_dl%3Dfyb_n_homepage%26hisfilter%3D1&logid=8309940529500911554&signature=4bce59041938b160b7c24423bde0b518&timestamp=1624535702')
 
    # 等待滑块出现
    WebDriverWait(driver, 10).until(lambda x: x.find_element_by_xpath('//div[contains(@class,"passMod_slide-btn")]'))
    yzm_button = driver.find_element_by_xpath('//div[contains(@class,"passMod_slide-btn")]')
    time.sleep(1)
    move_x = 100
 
    # 等待验证码出现
    WebDriverWait(driver, 10).until(lambda x: x.find_element_by_xpath('//img[contains(@class,"passMod_spin-background")]'))
    img_src = driver.find_element_by_xpath('//img[contains(@class,"passMod_spin-background")]').get_attribute('src')
 
    # 下载图片并转化为base64
    img_base64 = get_img(img_src)
    # 识别图片旋转角度
    move_x = shibie(img_base64)
    # 通过旋转角度 * 滑动系数 = 滑动距离
    move_x = move_x * 0.661
    # 开始滑动
    action = ActionChains(driver)
    action.click_and_hold(yzm_button).perform()  # 鼠标左键按下不放
    action.move_by_offset(move_x, 0).perform()
    action.release().perform()  # 释放鼠标
 
    time.sleep(2)
 
    # 第二次滑动
    # 等待滑块出现
    WebDriverWait(driver, 10).until(lambda x: x.find_element_by_xpath('//div[contains(@class,"passMod_slide-btn")]'))
    yzm_button = driver.find_element_by_xpath('//div[contains(@class,"passMod_slide-btn")]')
    time.sleep(1)
    move_x = 100
 
    # 等待验证码出现
    WebDriverWait(driver, 10).until(lambda x: x.find_element_by_xpath('//img[contains(@class,"passMod_spin-background")]'))
    img_src = driver.find_element_by_xpath('//img[contains(@class,"passMod_spin-background")]').get_attribute('src')
 
    # 下载图片并转化为base64
    img_base64 = get_img(img_src)
    # 识别图片旋转角度
    move_x = shibie(img_base64)
    # 通过旋转角度 * 滑动系数 = 滑动距离
    move_x = move_x * 0.661
    # 开始滑动
    action = ActionChains(driver)
    action.click_and_hold(yzm_button).perform()  # 鼠标左键按下不放
    action.move_by_offset(move_x, 0).perform()
    action.release().perform()  # 释放鼠标