Python+selenium破解拼圖驗證碼的腳本

更新時間：2022年02月07日 11:58:46 作者：我去熱飯

很多網(wǎng)站在登錄或者注冊時都會遇到拼圖驗證碼，這種拼圖驗證碼實際上是多個小碎片經(jīng)過重新組合成的一張整體。本文將和大家分享一個基于Python selenium的破解拼圖驗證碼的腳本，需要的可以參考一下

實現(xiàn)思路

很多網(wǎng)站都有拼圖驗證碼

1.首先要了解拼圖驗證碼的生成原理

2.制定破解計劃，考慮其可能性和成功率。

3.編寫腳本

很多網(wǎng)站的拼圖驗證碼都是直接借助第三方插件，也就是一類一種解法。

筆者遇到的這種拼圖驗證碼實際上是多個小碎片經(jīng)過重新組合成的一張整體，首先要在網(wǎng)站上抓取這種小碎片圖片并下載到本地

我們先捋一捋大體思路：

獲取所有碎片圖片----找出他們的排列順序邏輯-----找出他們中含有顏色深的真正位置的那個小碎塊的序號-----根據(jù)每塊碎片的寬度和上下和這個深色小塊的序號算出距離----用selenium向右移動滑塊到這個距離

核心代碼

先導(dǎo)入必要的包：

# -*- coding:utf-8 -*-
import colorsys
import urllib,os,uuid,re,time
from PIL import Image
from selenium.webdriver.common.action_chains import ActionChains
from selenium import webdriver

抓到并下載到本地

def create(locapath,fileName):
  filePath=locapath+'/'+fileName
  if not os.path.exists(filePath):
    file=open(filePath,'a+')
    file.close()
    return filePath
def downloadImg():
  list = openBrowser()
  for i in range(2):
    fileName = str(i) + '_test.jpg'
    urllib.urlretrieve(list[i], create('/Users/zijiawang/Downloads/', fileName))  #下載到本地
  time.sleep(3)
def openBrowser():
  global wzj
  wzj =  webdriver.Firefox()
  wzj.get('https://。。。。')
  image1_url = wzj.find_elements_by_class_name('gt_cut_bg_slice')[0].get_attribute('style')
  image1_url=image1_url[23:-38]
  image2_url = wzj.find_elements_by_class_name('gt_cut_fullbg_slice')[0].get_attribute('style')
  image2_url=image2_url[23:-38]
  return [image1_url,image2_url]

分割并獲取rgb顏色

ef getcolor(image):
  list = [[x,y] for x in range(26) for y in range(2)]
  listt = []
  colors = []
  for i in range(len(list)):
    l = list[i][0]*12+12
    w = list[i][1]*58+58
    listt.append([12*0.25+i/2*12, 58*0.25+((i+10)%2)*58])
    listt.append([12*0.25+i/2*12, 58*0.75+((i+10)%2)*58])
    listt.append([12*0.75+i/2*12, 58*0.25+((i+10)%2)*58])
    listt.append([12*0.75+i/2*12, 58*0.75+((i+10)%2)*58])
  for i in range(len(listt)):
    colors.append(image[listt[i][0],listt[i][1]])
  return colors

找出不同的小碎片序號

def getDeferent():
  deferent = []
  for i in range(208):
    if abs(color1[i][2] - color2[i][2]) >20:
      print u'B差值: ',color1[i][2]-color2[i][2]
      deferent.append(i)
      # print ([i,abs(color1[i][0] - color2[i][0]),abs(color1[i][1] - color2[i][1]),abs(color1[i][2] - color2[i][2])])
  return deferent
def getUPDOWN():
  deferent = getDeferent()
  xlist = []
  ylist = []
  for i in deferent:
    if (i / 4) % 2 == 0:
      xlist.append(i)
    else:
      ylist.append(i)
  uplist = []
  downlist = []
  for i in xlist:
    uplist.append(i / 4)
  for i in ylist:
    downlist.append(i / 4)
  for i in range(len(uplist)):
    uplist[i] /= 2
  for i in range(len(downlist)):
    downlist[i] /= 2
  func = lambda x, y: x if y in x else x + [y]
  up = reduce(func, [[], ] + uplist)
  func = lambda x, y: x if y in x else x + [y]
  down = reduce(func, [[], ] + downlist)
  return up,down

編寫好排列方法，會用的

def bubble(l):
    for index in range(len(l) - 1, 0 , -1):
        for two_index in range(index):
            if l[two_index] > l[two_index + 1]:
                l[two_index], l[two_index + 1] = l[two_index + 1], l[two_index]
    return l

最后的進行解析，拼裝源碼

def end():
  orders = getUPDOWN()
  orderx = orders[0]
  ordery = orders[1]
  print 'x:',orderx
  print 'y:',ordery
  orderX = [145, 157, 277, 265, 169, 181, 253, 241, 97, 109, 301, 289, 73, 85, 37, 25, 1, 13, 133, 121, 49, 61, 229,
            217, 193, 205]
  orderY = [157, 145, 265, 277, 181, 169, 241, 253, 109, 97, 289, 301, 85, 73, 25, 37, 13, 1, 121, 133, 61, 49, 217,
            229, 205, 193]
  print u'X:',orderX
  print u'Y:',orderY
  for x in range(26):
    orderX[x] -= 1
    orderX[x] /= 12
  for y in range(26):
    orderY[y] -= 1
    orderY[y] /= 12
  endX = []
  endY = []
  for i in range(len(orderX)):
    for j in range(len(orderx)):
      if orderx[j] == orderX[i]:
        endX.append(i)
        break
  for i in range(len(orderY)):
    for j in range(len(ordery)):
      if ordery[j] == orderY[i]:
        endY.append(i)
        break
  print 'endx:',endX
  print 'endy:',endY
  os.remove('/Users/zijiawang/Downloads/0_test.jpg')
  os.remove('/Users/zijiawang/Downloads/1_test.jpg')
  all = []
  all = endX+endY
  # print 'all:',all
  #系統(tǒng)最多可實驗5次
  func = lambda x, y: x if y in x else x + [y]
  all_old = reduce(func, [[], ] + all)
  # print 'all_old:',all_old
  all_end =bubble(all_old)
  print 'all_end:',all_end
  duandata = 1000
  for i in range(1,len(all_end)):
    if all_end[i]-1 in all_end:
      pass
    else:
      duandata = all_end[i]
      print u'斷點為:',duandata
  guess =[]
  if duandata != 1000:
    if endX != [] and endY!=[]:
      print u'斷點1'
      guess = [duandata,all_end[0]]
    elif endX ==[]:
      print u'斷點2'
      guess = [duandata, endY[0]]
    elif endY == []:
      print u'斷點3'
      guess = [duandata,  endX[0]]
  else:
      print u'無斷點'
      guess = [ all_end[0]]
  end_guess =[]
  print u'猜測的點:',guess
  for i in guess:
    end_guess.append(i * 12 - 38)
    end_guess.append(i * 12 - 28)
    end_guess.append(i * 12 - 20)
    end_guess.append(i * 12 - 17)
    end_guess.append(i * 12 - 15)
  print  u'猜測的位移量: ',end_guess
  return end_guess

主函數(shù)

if __name__ == '__main__':
  downloadImg()
  color1 = getcolor(Image.open('/Users/zijiawang/Downloads/0_test.jpg').load())
  color2 = getcolor(Image.open('/Users/zijiawang/Downloads/1_test.jpg').load())
  guess = end()
  for i in guess:
    print i,wzj.title
    try:
      ele = wzj.find_element_by_xpath('//div[@class="gt_slider_knob gt_show"]')
    except:
      ele = wzj.find_element_by_xpath('//div[@class="gt_slider_knob gt_show moving"]')
    ActionChains(wzj).click_and_hold(ele).perform()
    ActionChains(wzj).move_to_element_with_offset(ele, 0,i).perform()

因每個網(wǎng)站的都大同小異，這里不寫具體測試的網(wǎng)站了，以免對其造成壓力。此算法屬于高度定制的，應(yīng)用其他網(wǎng)站需要改些參數(shù)，但是代碼注釋較小。需要注釋，不明白的地方請留言哈。