Last active
February 12, 2023 03:47
-
-
Save TheSalarKhan/7c3d01ad13b0e7e5985a to your computer and use it in GitHub Desktop.
Background Averaging (Background Subtraction) in Python+OpenCV
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import numpy as np | |
import cv2 | |
class BackGroundSubtractor: | |
# When constructing background subtractor, we | |
# take in two arguments: | |
# 1) alpha: The background learning factor, its value should | |
# be between 0 and 1. The higher the value, the more quickly | |
# your program learns the changes in the background. Therefore, | |
# for a static background use a lower value, like 0.001. But if | |
# your background has moving trees and stuff, use a higher value, | |
# maybe start with 0.01. | |
# 2) firstFrame: This is the first frame from the video/webcam. | |
def __init__(self,alpha,firstFrame): | |
self.alpha = alpha | |
self.backGroundModel = firstFrame | |
def getForeground(self,frame): | |
# apply the background averaging formula: | |
# NEW_BACKGROUND = CURRENT_FRAME * ALPHA + OLD_BACKGROUND * (1 - APLHA) | |
self.backGroundModel = frame * self.alpha + self.backGroundModel * (1 - self.alpha) | |
# after the previous operation, the dtype of | |
# self.backGroundModel will be changed to a float type | |
# therefore we do not pass it to cv2.absdiff directly, | |
# instead we acquire a copy of it in the uint8 dtype | |
# and pass that to absdiff. | |
return cv2.absdiff(self.backGroundModel.astype(np.uint8),frame) | |
cam = cv2.VideoCapture(0) | |
# Just a simple function to perform | |
# some filtering before any further processing. | |
def denoise(frame): | |
frame = cv2.medianBlur(frame,5) | |
frame = cv2.GaussianBlur(frame,(5,5),0) | |
return frame | |
ret,frame = cam.read() | |
if ret is True: | |
backSubtractor = BackGroundSubtractor(0.01,denoise(frame)) | |
run = True | |
else: | |
run = False | |
while(run): | |
# Read a frame from the camera | |
ret,frame = cam.read() | |
# If the frame was properly read. | |
if ret is True: | |
# Show the filtered image | |
cv2.imshow('input',denoise(frame)) | |
# get the foreground | |
foreGround = backSubtractor.getForeground(denoise(frame)) | |
# Apply thresholding on the background and display the resulting mask | |
ret, mask = cv2.threshold(foreGround, 15, 255, cv2.THRESH_BINARY) | |
# Note: The mask is displayed as a RGB image, you can | |
# display a grayscale image by converting 'foreGround' to | |
# a grayscale before applying the threshold. | |
cv2.imshow('mask',mask) | |
key = cv2.waitKey(10) & 0xFF | |
else: | |
break | |
if key == 27: | |
break | |
cam.release() | |
cv2.destroyAllWindows() |
whats the difference between this and cv2.createBackgroundSubtractorMOG2()
@elhoussinetalab I think the difference is you can set the learning rate for static objects too, which you cant in cv2.createBackgroundSubtractorMOG2()
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Works great on Raspberry pi too (py2cv3)