OpenCV background subtraction via Python script

[AlexBtt]

Hi, i am currently trying to use the background subtractor MOG from OpenCV in my KNIME Workflow.
I have two images which i am subtracting. The code is working but now i am stuck on the implementation into KNIME. I tried using the PythonSourceNode but i don’t know how i can output the image for further use.
Here’s my code:

import cv2

backgroundSubtractor = cv2.bgsegm.createBackgroundSubtractorMOG()

# apply the algorithm for background images using learning rate > 0
for i in range(1, 16):
    bgImageFile = f"/data/bg{i}.jpg"
    bg = cv2.imread(bgImageFile)
    backgroundSubtractor.apply(bg, learningRate=0.5)

# apply the algorithm for detection image using learning rate 0
stillFrame = cv2.imread("/data/Test.jpg")
fgmask = backgroundSubtractor.apply(stillFrame, learningRate=0)

#output mask???

Thanks

[stelfrich]

Hi @AlexBtt,

@AlexBtt

I tried using the PythonSourceNode but i don’t know how i can output the image for further use.

There’s an extension called KNIME Image Processing - Python Extensions that changes the way how images are handled in a Python Script (1⇒1) node. Once you have this extension installed, you’d use KNIME Image Processing to load the two images and pass them to a Python Script node via the input table. The images are subsequently available as ndarrays in Python for further processing (see GitHub - knime-ip/knip-python-extensions: KNIME Image Processing Python Bindings for an example). You can then do your processing with OpenCV. But you have to make sure to wrap your output ndarray back into a KNIPImage and append it as a column:

# Convert output to KNIPImage
outputKnip = KNIPImage(output)
	
# Add to columns
column.append(outputKnip)

Best,
Stefan