Skip to content

chrisvoncsefalvay/processpathway

Repository files navigation

ProcessPathway

image

ProcessPathway is a nifty little tool that lets you play with image processing algorithms instead of wiring up your test bench all day. It is designed to feed a video camera input (a webcam, usually) through any functions that can digest and return a numpy matrix aka ndarray object (think OpenCV).

How to use

The example_threshold.py example, which - as the name suggests - thresholds the the webcam input, is a good example of what a typical ProcessPathway pipeline would look like.

First, import the LiveProcess object from processpathway. You might want to also import whatever processing functions you need.:

import cv2
from processpathway import LiveProcess

Then, you write your functions. The only constraint is that they should take and emit a numpy.ndarray image. Of course, that does not mean you can't code whatever side effects you need them to generate!:

def convert_to_grayscale(_frame):
    _frame = cv2.cvtColor(_frame, cv2.COLOR_BGR2GRAY)
    return _frame
def threshold(_frame):
    _, _frame = cv2.threshold(_frame, 128, 255, cv2.THRESH_BINARY)
    return _frame
def reconvert_to_bgr(_frame):
    _frame = cv2.cvtColor(_frame, cv2.COLOR_GRAY2BGR)
    return _frame

:

Upon call, attach LiveProcess to a variable, then bind the processing functions to it, in the order you want them to be performed! Then, bind a capture device and initialise it. Finally, start the loop.:

if __name__ == '__main__':
    processor = LiveProcess(fps=True)
    processor.bind_process(convert_to_grayscale, threshold, reconvert_to_bgr)
    processor.initialise_capture_device(0)
    processor.loop()

That's all, folks!

Built-in functionalities

  • FPS calculation (use fps=True when constructing your LiveProcess instance)

* Screenshotting (bound to the s key, on by default but can be disabled using screencap=False in the LiveProcess constructor * Automatic clean-up and closedown (using the esc key by default) * And lots more...

Todos

  • Tests! We desperately need some more tests.
  • Documentation! Might be autogeneratable as most of the functions are fairly extensively documented.
  • More examples! No better way to demonstrate the benefits of Processpathway.
  • Live histogram and other overlays.
  • I am not entirely happy with some functionalities yet. Among others, eventually I'd like the ability to inject observer blocks the way functions are injected, so that users would be able to have multiple output windows from the pipeline.

About

A Pythonic OpenCV (cv2) processing pathway.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages