數美驗證碼-空間推測-爬蟲

前言

因為要訓練識別驗證碼的模型，需要爬取源資料。
如果需要其他的型別，自行修改

爬取結果

這邊把圖片跟文字命名一樣的。

程式碼

import time
import json
import requests


def save_image_from_url(url, file_path):
    response = requests.get(url)
    with open(file_path, 'wb') as f:
        f.write(response.content)


if __name__ == '__main__':
    """
    @author: xXinG
    @dec:    爬取數美驗證碼空間 邏輯推理圖片、文字
    @time: 2023/4/26 19:50
    """

    execute_count = 100  # 爬取條數

    start_time = time.time()
    for i in range(execute_count):
        url = "https://captcha1.fengkongcloud.cn/ca/v1/register"

        querys = {"callback": "0", "channel": "DEFAULT",
                  "captchaUuid": "20240426194937QdanEmbtwJrRaJwbdG", "rversion": "1.0.4", "lang": "zh-cn",
                  "data": "{}", "model": "spatial_select", "sdkver": "1.1.3", "organization": "d6tpAY1oV0Kv5jRSgxQr",
                  "appId": "default"}

        response = requests.request("GET", url, params=querys)
        data_str = response.text[2:-1]  # 去掉字串開頭的'0('和結尾的')'
        data_dict = json.loads(data_str)  # 將字串轉換為字典

        if data_dict['code'] == 1100:
            url = data_dict['detail']['bg']  # 圖片url, 需要加上  https://castatic.fengkongcloud.cn/
            des = data_dict['detail']['order']  # 空間目標描述

            last_slash_index = url.rfind('/')
            jpg_start_index = url.find('.jpg')
            name = url[last_slash_index + 1:jpg_start_index]  # 圖片命名

            save_image_from_url('https://castatic.fengkongcloud.cn' + data_dict['detail']['bg'],
                                './result/img/' + name + '.jpg')

            with open("./result/label/" + name + '.txt', "w", encoding="utf-8") as f:
                f.write(des[0])
            print("完成 -> " + name)

    end_time = time.time()
    print("總共用時: " + str(end_time - start_time) + "秒")

這邊測試爬取100張

給大家看看訓練好的模型 哈哈