#coding=utf8
from selenium import webdriver
import pickle
import time
from PIL import Image
from PIL import ImageEnhance
from aip import AipOcr
def baidu_image_to_word(image_path):
""" 你的 APPID AK SK """
APP_ID = '14364432'
API_KEY = 'jgopMYaecGeGgaBr2EYWKNDZ'
SECRET_KEY = 'TnpKrHyyc3IgrGw2L5ZzKRiY9F2seCSk'
client = AipOcr(APP_ID, API_KEY, SECRET_KEY)
with open(image_path, 'rb') as fp:
image = fp.read()
""" 调用通用文字识别, 图片参数为本地图片 """
client.basicGeneral(image);
""" 如果有可选参数 """
options = {}
options["language_type"] = "ENG"
options["detect_direction"] = "true"#是否检测图像朝向，默认不检测，即：false
##options["detect_language"] = "true"#是否检测语言，默认不检测
options["probability"] = "true"#是否返回识别结果中每一行的置信度
""" 带参数调用通用文字识别, 图片参数为本地图片 """
res = client.basicAccurate(image, options)#通用文字识别(高精度版),普通版是client.basicGeneral(image, options);
##""" 调用通用文字识别, 图片参数为远程url图片 """
##url = "https//www.x.com/sample.jpg"
##client.basicGeneralUrl(url,options);
try:
guess = res['words_result'][0]['words']
probability = res['words_result'][0]['probability']['average']
except:
print("识别失败，将置信度归为0，文字为空")
guess = '';probability=0;
return guess,probability,res
browser = webdriver.Chrome()
browser.get('http://icrm.baidu.com/crm-portal/index.action')
browser.maximize_window()
time.sleep(2)
image = browser.find_element_by_xpath("//*[@id='imgCodeId']")
location = image.location# 获取验证码x,y轴坐标
size = image.size # 获取验证码的长宽
x_begin = int(location['x'])+5;
y_begin = int(location['y']);
my_with = size['width'] - 10
rangle = (x_begin,
y_begin,
x_begin+my_with,
y_begin+size['height']) # 写成我们需要截取的位置坐标
probability = 0;count = 0;count_max = 20;words_num = 0
##图片文字识别的置信度大于0.9的时候才说明识别得比较准确，否则刷新验证码重新识别
while (probability<0.9 or words_num!=4) and count if count>0: browser.find_element_by_xpath("//*[@id='imgCodeId']").click() time.sleep(2) count += 1 browser.save_screenshot('All.png') # 截取当前网页，该网页有我们需要的验证码 save_image = Image.open("All.png") # 打开截图 result = save_image.crop(rangle) result.save('result.jpg') ##增强图形识别率的处理 i2=Image.open(r"result.jpg") imgry = i2.convert('L') #图像加强，二值化，PIL中有九种不同模式。分别为1，L，P，RGB，RGBA，CMYK，YCbCr，I，F。L为灰度图像 sharpness =ImageEnhance.Contrast(imgry)#对比度增强 i3 = sharpness.enhance(3.0) #3.0为图像的饱和度 i3.save("result.png") ##连接api获取返回结果 guess,probability,res = baidu_image_to_word('result.png') #连接百度api words_num = len(guess) print('第%d次猜测验证码,猜测结果为%s,猜测验证码个数为%d,置信度平均值为%f'%(count,guess,words_num,probability)) if ' ' in guess: if len(guess.replace(' ',''))==4: guess = guess.replace(' ','') words_num = len(guess) else: probability = 0