Learning OpenCV 4 Computer Vision with Python 3 - Third Edition

Using a KNN background subtractor

By modifying just five lines of code in our MOG background subtraction script, we can use a different background subtraction algorithm, different morphology parameters, and a different video as input. Thanks to the high-level interface that OpenCV provides, even such simple changes enable us to successfully handle a wide variety of background subtraction tasks.

Just by replacing cv2.createBackgroundSubtractorMOG2 with cv2.createBackgroundSubtractorKNN, we can we use a background subtractor based on KNN clustering instead of MOG clustering:

bg_subtractor = cv2.createBackgroundSubtractorKNN(detectShadows=True)

Note that despite the change in algorithm, the detectShadows parameter is still supported. Additionally, ...