我们以这个网站为例: http://120.86.191.138/hbgs/zwgk/dirData.do?dirId=402881204e959150014e959f42f30014&subjectId=93e889f2501d3fe8015024305bdf0efc
往后点到第四页后会出现验证码

一.获取到背景图片和缺口图片

我们发现图片是base64格式通过API直接发送

二.识别缺口位置输出x距离

img_data = base64.b64decode(img_data)  # png
img_btn_data = base64.b64decode(img_btn_data)  # png
# 读取图片
img = Image.open(io.BytesIO(img_data))
img_btn = Image.open(io.BytesIO(img_btn_data))
# 识别缺口
left = identify_gap(img, img_btn, bg_params, gp_params)[0]
print(left)

identify_gap.py

import cv2
import numpy as np


def identify_gap(background, gap, bg_edge_list: list[int], gp_edge_list: list[int]):
    """
    获取缺口的X坐标

    使用OpenCV的边缘检测实现

    :param background: 背景图片对象
    :param gap: 缺口图片对象
        图片对象可以是cv2.imread('opencv_logo.jpg',0)
        也可以是PIL.Image.open('opencv_logo.jpg')
    :param bg_edge_list: 背景图片对比度调整参数[50, 100] 50为最小对比度，100为最大对比度
    :param gp_edge_list:缺口图片对比度调整参数[50, 100] 50为最小对比度，100为最大对比度
    :return:
    """

    background_image = background  # 背景图片
    background_image = np.array(background_image)
    gap_image = gap  # 缺口图片
    gap_image = np.array(gap_image)
    # 识别图片边缘
    background_edge = cv2.Canny(background_image, bg_edge_list[0], bg_edge_list[1])
    gap_edge = cv2.Canny(gap_image, gp_edge_list[0], gp_edge_list[1])

    # 转换图片格式
    background_picture = cv2.cvtColor(background_edge, cv2.COLOR_GRAY2RGB)

    gap_picture = cv2.cvtColor(gap_edge, cv2.COLOR_GRAY2RGB)

    res = cv2.matchTemplate(background_picture, gap_picture, cv2.TM_CCOEFF_NORMED)
    min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(res)  # type: ignore

    return max_loc

3.完整代码

import base64
import io
import re
import requests
from PIL import Image
from packages.scrapy_spiders.utils.captcha.compare_image import identify_gap
def getlist_cookie(yzm_res):
    """
    获取验证码返回的set-cookie并返回
    :param response:
    :return dic:
    """
    extracted_cookies = {}
    cookies = yzm_res.headers["Set-Cookie"].split(",")
    # 解析Set-Cookie头部中的cookies
    for cookie in cookies:
        parts = cookie.split(";")[0].split("=")
        if len(parts) == 2:
            key, value = parts
            extracted_cookies[key.strip()] = value.strip()
    return extracted_cookies
def generate_random_string():
    htmlurl='http://120.86.191.138/hbgs/verify/index.do'
    response = requests.get(htmlurl).text
    # 编写正则表达式来匹配verify后的值，注意这里我们使用了转义字符来匹配冒号和单引号
    pattern = r"verify\s*:\s*'([^']*)'"

    # 使用re.search查找匹配项
    match = re.search(pattern, response)

    # 如果找到匹配项，则提取括号内的内容（即需要的值）
    if match:
        value = match.group(1)
        print("提取的verify值为:", value)
    else:
        print("未找到匹配的verify值")
        return None

    return value

ssst=generate_random_string()

yzm_url=f'http://120.86.191.138/hbgs/verify/get_img_verify.do?verify={ssst}'
res=requests.get(yzm_url)
cookies = getlist_cookie(res)
res=res.json()
bg_params = [100, 200]
gp_params = [100, 200]
img_data=res["oriCopyImage"]
img_btn_data=res["newImage"]
img_data = base64.b64decode(img_data)  # jpg
img_btn_data = base64.b64decode(img_btn_data)  # png
# 读取图片
img = Image.open(io.BytesIO(img_data))
img_btn = Image.open(io.BytesIO(img_btn_data))
# 识别缺口
left = identify_gap(img, img_btn, bg_params, gp_params)[0]
print(left)

dataurl='http://120.86.191.138/hbgs/verify/check.do'
header={
'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8'
}
formdata=f'verify={ssst}&IY={res["Y"]}&X={left}&Y={res["Y"]}'

res2=requests.post(dataurl, formdata,cookies=cookies,headers=header).json()
print(res2)
itemurl='http://120.86.191.138/hbgs/zwgk/item.do'
formdata=f'page=6&rows=20&HBTB_XH=&HBTB_XH_END=&HBTB_XMMC=&HBTB_SPWH=&HBTB_TXDZ=&HBTB_GSSJ=&HBTB_GSSJ_END=&dirId=402881204e959150014e95bb85b5010f&subjectId=93e889f2501d3fe8015024305bdf0efc&backPage=&vcode={res2["data"]}'

res=requests.post(itemurl, formdata)
print(res.text)