我想从浏览器获取验证码图片。我已经获得了该图片的URL，但是这张图片每次更新时都会更改（URL不变）。
是否有任何解决方案可以从浏览器中获取图片（例如“保存图像”按钮）？
另一方面，我认为以下方法应该有效：
动态验证码链接-链接
问题已通过截屏解决：
browser.save_screenshot('screenshot.png')
img = browser.find_element_by_xpath('//*[@id="cryptogram"]')
loc = img.location
image = cv.LoadImage('screenshot.png', True)
out = cv.CreateImage((150,60), image.depth, 3)
cv.SetImageROI(image, (loc['x'],loc['y'],150,60))
cv.Resize(image, out)
cv.SaveImage('out.jpg', out)
感谢
这里是一个完整的例子（使用谷歌的reCAPTCHA作为目标）：
import urllib
from selenium import webdriver
driver = webdriver.Firefox()
driver.get('http://www.google.com/recaptcha/demo/recaptcha')
# get the image source
img = driver.find_element_by_xpath('//div[@id="recaptcha_image"]/img')
src = img.get_attribute('src')
# download the image
urllib.urlretrieve(src, "captcha.png")
driver.close()
更新：
动态生成的图像问题在于每次请求时都会生成新的图像。在这种情况下，您有几个选项：
截屏
from selenium import webdriver
driver = webdriver.Firefox()
driver.get('https://moscowsg.megafon.ru/ps/scc/php/cryptographp.php?PHPSESSID=mfc540jkbeme81qjvh5t0v0bnjdr7oc6&ref=114&w=150')
driver.save_screenshot("screenshot.png")
driver.close()
模拟右键点击+“另存为”。有关更多信息，请参见此线程。
可以保存整个页面的截图然后从中裁剪图像，但是您还可以使用“webdriver”中的“find”方法来定位要保存的图像，并像下面这样编写“screenshot_as_png”属性：
from selenium import webdriver
driver = webdriver.Firefox()
driver.get('https://www.webpagetest.org/')
with open('filename.png', 'wb') as file:
file.write(driver.find_element_by_xpath('/html/body/div[1]/div[5]/div[2]/table[1]/tbody/tr/td[1]/a/div').screenshot_as_png)
selenium.common.exceptions.WebDriverException: Message: unknown command: session/5734e4b0f8d6171317af42ddf0979562/element/0.9586500909174849-1/screenshot。 - Jortegasave_screenshot的问题在于我们不能保存图像的原始质量，并且无法还原图像中的 alpha 通道。因此，我提出了另一种解决方案。以下是一个完整的示例，使用了 @codam_hsmits 建议的 selenium-wire 库。可以通过ChromeDriver下载图像。from seleniumwire import webdriver # Import from seleniumwire
from urllib.parse import urlparse
import os
from mimetypes import guess_extension
import time
import datetime
def download_assets(requests,
asset_dir="temp",
default_fname="unnamed",
skip_domains=["facebook", "google", "yahoo", "agkn", "2mdn"],
exts=[".png", ".jpeg", ".jpg", ".svg", ".gif", ".pdf", ".bmp", ".webp", ".ico"],
append_ext=False):
asset_list = {}
for req_idx, request in enumerate(requests):
# request.headers
# request.response.body is the raw response body in bytes
if request is None or request.response is None or request.response.headers is None or 'Content-Type' not in request.response.headers:
continue
ext = guess_extension(request.response.headers['Content-Type'].split(';')[0].strip())
if ext is None or ext == "" or ext not in exts:
#Don't know the file extention, or not in the whitelist
continue
parsed_url = urlparse(request.url)
skip = False
for d in skip_domains:
if d in parsed_url.netloc:
skip = True
break
if skip:
continue
frelpath = parsed_url.path.strip()
if frelpath == "":
timestamp = str(datetime.datetime.now().replace(microsecond=0).isoformat())
frelpath = f"{default_fname}_{req_idx}_{timestamp}{ext}"
elif frelpath.endswith("\\") or frelpath.endswith("/"):
timestamp = str(datetime.datetime.now().replace(microsecond=0).isoformat())
frelpath = frelpath + f"{default_fname}_{req_idx}_{timestamp}{ext}"
elif append_ext and not frelpath.endswith(ext):
frelpath = frelpath + f"_{default_fname}{ext}" #Missing file extension but may not be a problem
if frelpath.startswith("\\") or frelpath.startswith("/"):
frelpath = frelpath[1:]
fpath = os.path.join(asset_dir, parsed_url.netloc, frelpath)
if os.path.isfile(fpath):
continue
os.makedirs(os.path.dirname(fpath), exist_ok=True)
print(f"Downloading {request.url} to {fpath}")
asset_list[fpath] = request.url
try:
with open(fpath, "wb") as file:
file.write(request.response.body)
except:
print(f"Cannot download {request.url} to {fpath}")
return asset_list
让我们从谷歌主页下载一些图片到temp文件夹中。
# Create a new instance of the Chrome/Firefox driver
driver = webdriver.Chrome()
# Go to the Google home page
driver.get('https://www.google.com')
# Download content to temp folder
asset_dir = "temp"
while True:
# Please browser the internet, it will collect the images for every second
time.sleep(1)
download_assets(driver.requests, asset_dir=asset_dir)
driver.close()
为了保持与时俱进，这里介绍一个2020年的解决方案，使用seleniumwire库来获取浏览器发出的请求。您可以按照以下方式轻松使用：
from seleniumwire import webdriver
# Sometimes, selenium randomly crashed when using seleniumwire, these options fixed that.
# Probably has to do with how it proxies everything.
options.add_argument('--ignore-certificate-errors')
options.add_argument('--ignore-ssl-errors')
driver = webdriver.Chrome(chrome_options=options)
driver.get("https://google.com")
for request in driver.requests:
# request.path
# request.method
# request.headers
# request.response is the response instance
# request.response.body is the raw response body in bytes
# if you are using it for a ton of requests, make sure to clear them:
del driver.requests
现在，为什么需要这个？比如说用于ReCaptcha绕过或绕过像Incapsula这样的东西。请自行决定使用风险。
您可以使用JS下载图像而不会失去质量：
from io import BytesIO
from PIL import Image
from base64 import b64decode
driver.get(url)
# Create a canvas, set it's width and height equal to image's
# Write image to canvas, translate to base64
# Remove metadata prefix
b64img = driver.execute_script(r'''
var img = document.getElementsByTagName("img")[0];
var canvas = document.createElement("canvas");
canvas.width = img.width;
canvas.height = img.height;
var ctx = canvas.getContext("2d");
ctx.drawImage(img, 0, 0);
var dataURL = canvas.toDataURL("image/png");
return dataURL.replace(/^data:image\/(png|jpg);base64,/, "");
''')
# Decode from base64, translate to bytes and write to PIL image
img = Image.open(BytesIO(b64decode(b64img)))
from selenium import webdriver
#set chromedriver.exe path
driver = webdriver.Chrome()
driver.implicitly_wait(0.5)
#maximize browser
driver.maximize_window()
#launch URL
driver.get("name of webpage from where you want to download image");
#open file in write and binary mode
with open('Logo.png', 'wb') as file:
#identify image to be captured
l = driver.find_element_by_xpath('//*[@alt="name in the alt of image"]')
#write file
file.write(l.screenshot_as_png)
#close browser
driver.quit()
如果您需要授权才能下载图像，我发现最好的解决方案是结合selenium和selenium-requests：
response = driver.request("GET", image_url, stream=True)
response.raise_for_status()
with open(path, 'wb') as f:
response.raw.decode_content = True
shutil.copyfileobj(response.raw, f)
from seleniumwire import webdriver
seleniumwire_options = {
"disable_encoding": True,
}
driver = webdriver.Chrome(options=options, seleniumwire_options=seleniumwire_options)
driver.get(url) # direct url to image
request = driver.last_request
file = url.split("/")[-1]
with open(file, "wb") as f:
f.write(request.response.body)
这里是。
BeautifulSoup提取图像的宽度和高度driver.set_window_size设置正确的当前窗口大小，并使用driver.save_screenshot截屏from bs4 import BeautifulSoup
from selenium import webdriver
import os
from urllib.parse import urlparse
url = 'https://image.rakuten.co.jp/azu-kobe/cabinet/hair1/hb-30-pp1.jpg'
filename = os.path.basename(urlparse(url).path)
filename_png = os.path.splitext(filename)[0] + '.png' # change file extension to .png
opts = webdriver.ChromeOptions()
opts.headless = True
driver = webdriver.Chrome(options=opts)
driver.get(url)
# Get the width and height of the image
soup = BeautifulSoup(driver.page_source, 'lxml')
width = soup.find('img')['width']
height = soup.find('img')['height']
driver.set_window_size(width, height) # driver.set_window_size(int(width), int(height))
driver.save_screenshot(filename_png)
这同样适用于Google的图片格式WebP。