Take a screenshot of a web page in Java

java screenshot

51,875

Solution 1

To build on two of the answers above:

Rendering the HTML in Java then saving to an image - A few Java based HTML renders exist, all with different sets of drawbacks. The most common is the one built in. This is quite simple and can only render fairly basic HTML. The most intresting I know of is The Flying Saucer Project. This can render fairly complex XHTML but you will have to convert HTML before you can use it (JTindy may be able to help here). Taking a Swing component and creating a image is fairly simple, you just pass an BufferedImages graphics object and pass it to the Swing components paint method. Then splat that out with ImageIO.
A big advantage to this would be that the renderer would be headless. The disadvantage is that it would not be a perfect rendering and it would lack any plugins.

The second option requires you to start a web browser, work out where it is and then take a screen shot. Optionally you may also wish to strip out all the Firefox/IE/Opera/etc menus leaving you with just image. To get the dimensions of the web browser the simplest option would be to start it full screen. The other option would be to use something like JDICs browser component to include it as part of the Java application. It would then be able to specify where the HTML is being rendered on screen and then simply use Robot to create a screen shot of that area.
The big advantage to this is that it will give a perfect rendering (for a given browser). The two disadvantages is that it would require native code (or at least using a native component) and it could not be headless¹.

1) You could use a virtual frame buffer. But that is outside Java.

Solution 2

I had best results with Selenium Webdriver using a VirtualFramebuffer and Firefox Binary. This is tested under ubuntu. You need to have xvfb and firefox installed

First install firefox and virtual framebuffer:

aptitude install xvfb firefox

Compile and run this class, open /tmp/screenshot.png afterwards

import java.io.File;
import java.io.IOException;

import org.apache.commons.io.FileUtils;
import org.openqa.selenium.OutputType;
import org.openqa.selenium.TakesScreenshot;
import org.openqa.selenium.WebDriver;
import org.openqa.selenium.firefox.FirefoxBinary;
import org.openqa.selenium.firefox.FirefoxDriver;

public class CaptureScreenshotTest
{
    private static int      DISPLAY_NUMBER  = 99;
    private static String   XVFB            = "/usr/bin/Xvfb";
    private static String   XVFB_COMMAND    = XVFB + " :" + DISPLAY_NUMBER;
    private static String   URL             = "http://www.google.com/";
    private static String   RESULT_FILENAME = "/tmp/screenshot.png";

    public static void main ( String[] args ) throws IOException
    {
        Process p = Runtime.getRuntime().exec(XVFB_COMMAND);
        FirefoxBinary firefox = new FirefoxBinary();
        firefox.setEnvironmentProperty("DISPLAY", ":" + DISPLAY_NUMBER);
        WebDriver driver = new FirefoxDriver(firefox, null);
        driver.get(URL);
        File scrFile = ( (TakesScreenshot) driver ).getScreenshotAs(OutputType.FILE);
        FileUtils.copyFile(scrFile, new File(RESULT_FILENAME));
        driver.close();
        p.destroy();
    }
}

Solution 3

It's not Java, but after coming here, it's what I ended up using, so I think it's worth a mention. With PhantomJs you can run a headless version of webkit, and then access the functionality through a built in mongoose webserver, which can handle requests for screencaptures, and store them locally. You can use Java to the make the request, and the response can have the url to the image on the server, so you could grab that too-

Solution 4

use selenium-rc

Solution 5

You can use the createScreenCapture method in awt.Robot. This method allows you to specify which portion of the screen to capture. So, you would still need to determine the coordinates of the window that contains the web page that you want to capture.

View more solutions

51,875