从图像中切出或移除进入蒙版黑色像素的像素

Question

I'm trying to cut out two images with OpenCV using Python.

The idea is to use an image like this one:

As a mask. Note: this is only an example.

So if I subtract this mask to another image, I will get an image with the original pixels only in the same location as the white rectangle in the mask. So, the idea is to remove the pixels from the first image that run into the same location that the black ones in the mask.

How can I do it?

Maybe I can do an AND operation between the two images (the mask will have only 0 (for black) and 1 (for white pixels).

Answer 1

If what you are asking is how to make the image transparent where the mask is black and opaque where the mask is white, then just add the mask to the alpha channel of the input. Here is how to do that in Python/OpenCV. (Revised per comment by Mark Setchell)

Input:

Mask:

import cv2
import numpy as np

# load image
img = cv2.imread('lena.png')

# load mask as grayscale
mask = cv2.imread('rect_mask.png', cv2.COLOR_BGR2GRAY)

# put mask into alpha channel of image
#result = img.copy()
#result = cv2.cvtColor(result, cv2.COLOR_BGR2BGRA)
#result[:, :, 3] = mask
result = np.dstack((img, mask))

# save resulting masked image
cv2.imwrite('lena_masked.png', result)

Result:

Answer 2

This is how I did it.

Input arrays have a shape of (240, 240) and output array has a shape of (240, 240) .

I will mask the indices on image array with the indices in mask array which have a value of 0.

def cut_out(image, mask):
    if type(image) != np.ndarray:
        raise TypeError("image must be a Numpy array")
    elif type(mask) != np.ndarray:
        raise TypeError("mask must be a Numpy array")
    elif image.shape != mask.shape:
        raise ValueError("image and mask must have the same shape")

    return np.where(mask==0, 0, image)