CODEWITHSUNDEEP
pythonopencvbackground-subtraction
slow_one
slow_one

Reputation: 113

'Module' object has no attribute 'BackgroundSubtractorMOG'

I'm trying to use the default functions to do background subtraction of a video file.
I'm using Python 2.7 and OpenCV.
I receive this error when using the 'BackgroundSubtractorMOG' module: 

'module' object has no attribute 'BackgroundSubtractorMOG'

Now, if I try using the createBackgroundSubtractorMOG() module, I receive the same error.
If I change my code to:

bkgnd = cv2.bgsegm.BackgroundSubtractorMOG()

I get an error later on that tries to tell me I'm using OpenCV 3.1 (although I'm fairly certain I'm not).

 img_sub_gray_image = cv2.cvtColor(img_sub, cv2.COLOR_BGR2GRAY) cv2.error:

/home/odroid/opencv-3.1.0/modules/imgproc/src/color.cpp:8000: error: (-215) scn == 3 || scn == 4 in function cvtColor

Here is the segment of code that is erroring out:

bkgnd = cv2.bgsegm.BackgroundSubtractorMOG() 
cap = cv2.VideoCapture('video.mp4')

while(True): 
    ret, frame = cap.read()

    #background subtraction
    img_sub = bkgnd.apply(frame)

    #convert to grayscale
    img_sub_gray_image = cv2.cvtColor(img_sub, cv2.COLOR_BGR2GRAY)
    #thresholding, forcing to binary image
    ret,threshold1 = cv2.threshold(img_sub_gray_image, LOWER_BOUND, UPPER_BOUND, cv2.THRESH_BINARY)

Any ideas?

Upvotes: 2

Views: 10282

Answers (3)

Tolga Sahin
Tolga Sahin

Reputation: 347

if you are using opencv 4 you should use MOG2 with history, varThreshold, detectShadows parameters

cv2.createBackgroundSubtractorMOG2(history=100,varThreshold=50,detectShadows=True)

Upvotes: 0

DataTales
DataTales

Reputation: 19

This works well..

fgbg = cv2.BackgroundSubtractorMOG2()

P.S: I am using python 3.5 and cv2 version 3.3.1

Upvotes: 0

eyllanesc
eyllanesc

Reputation: 244003

In opencv 3 they have changed some names of the functions, you must change:

bkgnd = cv2.bgsegm.BackgroundSubtractorMOG()

to:

cv2.bgsegm.createBackgroundSubtractorMOG()

Another observation is that the result of the apply function is a binary image, so it is not necessary to do the conversion from RGB to gray. You do not need to use the command: cv2.cvtColor(img_sub, cv2.COLOR_BGR2GRAY)

Upvotes: 3

Related Questions