Consecutive Image Masking in Python

Question

I am working on a project to mask consecutive images at different positions to study the movement of a single fly. The final image should look like this: correct masking of image. However, as the code continues to run, the images start to turn out like this: Incorrect image masking, which is not what I want because I only want one visible region (circle containing a single fly) per image rather than three. What do I need to do to my code so that only one circle is present in each image at different positions rather than there being multiple circles in an image?

`List = [
    [180, 353, 85, str("/media/pi/T3 1TB 2Ch xFer/Ramzy Masked /Arena2")],
    [180, 542, 85, str("/media/pi/T3 1TB 2Ch xFer/Ramzy Masked /Arena3")],
    [180, 728, 85, str("/media/pi/T3 1TB 2Ch xFer/Ramzy Masked /Arena4")],
]

def load_images_from_folder(folder):
    images = []
    for filename in os.listdir(folder):
        img = cv2.imread(os.path.join(folder,filename))
        mask = np.zeros(img.shape[:2], dtype="uint8")
    
        for row in List:
            cv2.circle(mask, (row[0], row[1]), row[2], 255, -1)
            masked = cv2.bitwise_and(img, img, mask=mask)
            cv2.imwrite(os.path.join(row[3], f'{filename}'),  masked)

load_images_from_folder("/media/pi/T3 1TB 2Ch xFer/Ramzy Research 
Photos/8.17.21")
    
cv2.waitKey(0)
cv2.destroyAllWindows()

`

Snuh · Accepted Answer · 2021-08-20T16:53:02.243

0

I am not overly familiar with the cv2 library but I believe that this line:

cv2.circle(mask, (row[0], row[1]), row[2], 255, -1)

Is adding a new circle onto your mask at every loop. There are a bunch of ways to fix this, but you could begin with moving this line to just above the cv2.circle call in the loop like such:

def load_images_from_folder(folder):
    images = []
    for filename in os.listdir(folder):
        img = cv2.imread(os.path.join(folder,filename))
    
        for row in List:
            mask = np.zeros(img.shape[:2], dtype="uint8")
            cv2.circle(mask, (row[0], row[1]), row[2], 255, -1)
            masked = cv2.bitwise_and(img, img, mask=mask)
            cv2.imwrite(os.path.join(row[3], f'{filename}'),  masked)

This will create a brand new mask on each loop iteration.

edited Aug 20 '21 at 16:53

answered Aug 20 '21 at 03:13

Snuh

310
2
10

When I do that, it says that "local variable 'row' referenced before assignment". I think that since the for loop defines what a row is, that line has to be where it was originally. Any ideas about what to do next? – Ramzy Issa Aug 20 '21 at 16:31
@RamzyIssa sorry, my comment wasn't clear enough. I've edited my comment with the whole function, moving the mask variable into the inner loop. – Snuh Aug 20 '21 at 16:54

Consecutive Image Masking in Python

1 Answers1