RPA.Images

Images is a library for general image manipulation. For image-based desktop automation, use the RPA.Desktop library.

Coordinates

The coordinates used in the library are pairs of x and y values that represent pixels. The upper left corner of the image or screen is (0, 0). The x-coordinate increases towards the right, and the y-coordinate increases towards the bottom.

Regions are represented as tuples of (left, top, right, bottom). For example, a 400 by 200-pixel region in the upper left corner would be (0, 0, 400, 200).

Template matching

Template matching refers to an operation where the (potential) location of a smaller image is searched from a larger image. It can be used for verifying certain conditions or locating UI elements for desktop or web automation.

Requirements

The default installation depends on Pillow library, which is used for general image manipulation operations.

For more robust and faster template matching, the library can use a combination of NumPy and OpenCV. They can be installed by opting in to the recognition dependency:

pip install rpaframework rpaframework-recognition

Examples

Robot Framework

The Images library can be imported and used directly in Robot Framework, for instance, for capturing screenshots or verifying something on the screen.

Desktop automation based on images should be done using the corresponding desktop library, i.e. RPA.Desktop.

*** Settings *** Library RPA.Images *** Keywords *** Should show success [Documentation] Raises ImageNotFoundError if success image is not on screen Find template on screen ${CURDIR}${/}success.png Save screenshot to results [Documentation] Saves screenshot of desktop with unique name ${timestamp}= Get current date result_format=%H%M%S Take screenshot filename=${OUTPUT_DIR}${/}desktop_${timestamp}.png

Python

from RPA.Images import Images def draw_matches_on_image(source, template): matches = lib.find_template_in_image(source, template) for match in matches: lib.show_region_in_image(source, match) source.save("matches.png")