Screenshot Functions

PyAutoGUI can take screenshots, save them to files, and locate images within the screen. This is useful if you have a small image of, say, a button that needs to be clicked and want to locate it on the screen. These features are provided by the PyScreeze module, which is installed with PyAutoGUI.

Screenshot functionality requires the Pillow module. OS X uses the screencapture command, which comes with the operating system. Linux uses the scrot command, which can be installed by running sudo apt-get install scrot.

The screenshot() Function

Calling screenshot() will return an Image object (see the Pillow or PIL module documentation for details). Passing a string of a filename will save the screenshot to a file as well as return it as an Image object.

>>> import pyautogui
>>> im1 = pyautogui.screenshot()
>>> im2 = pyautogui.screenshot('my_screenshot.png')

On a 1920 x 1080 screen, the screenshot() function takes roughly 100 milliseconds - it’s not fast but it’s not slow.

There is also an optional region keyword argument, if you do not want a screenshot of the entire screen. You can pass a four-integer tuple of the left, top, width, and height of the region to capture:

>>> import pyautogui
>>> im = pyautogui.screenshot(region=(0,0, 300, 400))

The Locate Functions

NOTE: As of version 0.9.41, if the locate functions can’t find the provided image, they’ll raise ImageNotFoundException instead of returning None.

You can visually locate something on the screen if you have an image file of it. For example, say the calculator app was running on your computer and looked like this:

_images/calculator.png

You can’t call the moveTo() and click() functions if you don’t know the exact screen coordinates of where the calculator buttons are. The calculator can appear in a slightly different place each time it is launched, causing you to re-find the coordinates each time. However, if you have an image of the button, such as the image of the 7 button:

_images/calc7key.png

… you can call the locateOnScreen('calc7key.png') function to get the screen coordinates. The return value is a 4-integer tuple: (left, top, width, height). This tuple can be passed to center() to get the X and Y coordinates at the center of this region. If the image can’t be found on the screen, locateOnScreen() raises ImageNotFoundException.

>>> import pyautogui
>>> button7location = pyautogui.locateOnScreen('calc7key.png')
>>> button7location
Box(left=1416, top=562, width=50, height=41)
>>> button7location[0]
1416
>>> button7location.left
1416
>>> button7point = pyautogui.center(button7location)
>>> button7point
Point(x=1441, y=582)
>>> button7point[0]
1441
>>> button7point.x
1441
>>> button7x, button7y = button7point
>>> pyautogui.click(button7x, button7y)  # clicks the center of where the 7 button was found
>>> pyautogui.click('calc7key.png') # a shortcut version to click on the center of where the 7 button was found

The optional confidence keyword argument specifies the accuracy with which the function should locate the image on screen. This is helpful in case the function is not able to locate an image due to negligible pixel differences:

>>> import pyautogui
>>> button7location = pyautogui.locateOnScreen('calc7key.png', confidence=0.9)
>>> button7location
Box(left=1416, top=562, width=50, height=41)

Note: You need to have OpenCV installed for the confidence keyword to work.

The locateCenterOnScreen() function combines locateOnScreen() and center():

>>> import pyautogui
>>> x, y = pyautogui.locateCenterOnScreen('calc7key.png')
>>> pyautogui.click(x, y)

On a 1920 x 1080 screen, the locate function calls take about 1 or 2 seconds. This may be too slow for action video games, but works for most purposes and applications.

There are several “locate” functions. They all start looking at the top-left corner of the screen (or image) and look to the right and then down. The arguments can either be a

  • locateOnScreen(image, grayscale=False) - Returns (left, top, width, height) coordinate of first found instance of the image on the screen. Raises ImageNotFoundException if not found on the screen.
  • locateCenterOnScreen(image, grayscale=False) - Returns (x, y) coordinates of the center of the first found instance of the image on the screen. Raises ImageNotFoundException if not found on the screen.
  • locateAllOnScreen(image, grayscale=False) - Returns a generator that yields (left, top, width, height) tuples for where the image is found on the screen.
  • locate(needleImage, haystackImage, grayscale=False) - Returns (left, top, width, height) coordinate of first found instance of needleImage in haystackImage. Raises ImageNotFoundException if not found on the screen.
  • locateAll(needleImage, haystackImage, grayscale=False) - Returns a generator that yields (left, top, width, height) tuples for where needleImage is found in haystackImage.

The “locate all” functions can be used in for loops or passed to list():

>>> import pyautogui
>>> for pos in pyautogui.locateAllOnScreen('someButton.png')
...   print(pos)
...
(1101, 252, 50, 50)
(59, 481, 50, 50)
(1395, 640, 50, 50)
(1838, 676, 50, 50)
>>> list(pyautogui.locateAllOnScreen('someButton.png'))
[(1101, 252, 50, 50), (59, 481, 50, 50), (1395, 640, 50, 50), (1838, 676, 50, 50)]

These “locate” functions are fairly expensive; they can take a full second to run. The best way to speed them up is to pass a region argument (a 4-integer tuple of (left, top, width, height)) to only search a smaller region of the screen instead of the full screen:

>>> import pyautogui
>>> pyautogui.locateOnScreen('someButton.png', region=(0,0, 300, 400))

Grayscale Matching

Optionally, you can pass grayscale=True to the locate functions to give a slight speedup (about 30%-ish). This desaturates the color from the images and screenshots, speeding up the locating but potentially causing false-positive matches.

>>> import pyautogui
>>> button7location = pyautogui.locateOnScreen('calc7key.png', grayscale=True)
>>> button7location
(1416, 562, 50, 41)

Pixel Matching

To obtain the RGB color of a pixel in a screenshot, use the Image object’s getpixel() method:

>>> import pyautogui
>>> im = pyautogui.screenshot()
>>> im.getpixel((100, 200))
(130, 135, 144)

Or as a single function, call the pixel() PyAutoGUI function, which is a wrapper for the previous calls:

>>> import pyautogui
>>> pix = pyautogui.pixel(100, 200)
>>> pix
RGB(red=130, green=135, blue=144)
>>> pix[0]
130
>>> pix.red
130

If you just need to verify that a single pixel matches a given pixel, call the pixelMatchesColor() function, passing it the X coordinate, Y coordinate, and RGB tuple of the color it represents:

>>> import pyautogui
>>> pyautogui.pixelMatchesColor(100, 200, (130, 135, 144))
True
>>> pyautogui.pixelMatchesColor(100, 200, (0, 0, 0))
False

The optional tolerance keyword argument specifies how much each of the red, green, and blue values can vary while still matching:

>>> import pyautogui
>>> pyautogui.pixelMatchesColor(100, 200, (130, 135, 144))
True
>>> pyautogui.pixelMatchesColor(100, 200, (140, 125, 134))
False
>>> pyautogui.pixelMatchesColor(100, 200, (140, 125, 134), tolerance=10)
True