Create buffer zone within a Numpy array

Question

I have a binary image as follows:

data = np.array([[0, 0 , 0 , 0 , 0 , 0 , 0 , 0],
                 [0, 0 , 0 , 0 , 0 , 0 , 0 , 0],
                 [0, 0 , 1 , 1 , 1 , 1 , 0 , 0],
                 [0, 0 , 1 , 1 , 1 , 1 , 0 , 0],
                 [0, 0 , 1 , 1 , 1 , 1 , 0 , 0],
                 [0, 0 , 0 , 0 , 0 , 0 , 0 , 0],
                 [0, 0 , 0 , 0 , 0 , 0 , 0 , 0]])

For pixels having 1s values, I want to make buffer zone of two pixels with value 1s surrounded in every four directions. The expected result would be:

result=np.array([[1, 1 , 1 , 1 , 1 , 1 , 1 , 1],
                 [1, 1 , 1 , 1 , 1 , 1 , 1 , 1],
                 [1, 1 , 1 , 1 , 1 , 1 , 1 , 1],
                 [1, 1 , 1 , 1 , 1 , 1 , 1 , 1],
                 [1, 1 , 1 , 1 , 1 , 1 , 1 , 1],
                 [1, 1 , 1 , 1 , 1 , 1 , 1 , 1],
                 [1, 1 , 1 , 1 , 1 , 1 , 1 , 1]])

How can I do it?

Answer 1

If you only have ones and zeros on the input and output array, you can do it with a 2D convolution, which is simple and works.

from scipy.signal import convolve2d

data = np.array([[0, 0 , 0 , 0 , 0 , 0 , 0 , 0],
                 [0, 0 , 0 , 0 , 0 , 0 , 0 , 0],
                 [0, 0 , 1 , 1 , 1 , 1 , 0 , 0],
                 [0, 0 , 1 , 1 , 1 , 1 , 0 , 0],
                 [0, 0 , 1 , 1 , 1 , 1 , 0 , 0],
                 [0, 0 , 0 , 0 , 0 , 0 , 0 , 0],
                 [0, 0 , 0 , 0 , 0 , 0 , 0 , 0]])

# the kernel doesn't need to be ones, it just needs to be positive and
# non-zero.
kernel = np.ones((5, 5))

result = np.int64(convolve2d(data, kernel, mode='same') > 0)

Which gives you the output you want. You need to define what you want to happen at the edges - in this version, the output array is the same size as the input array.

It might be possible you can do something faster if you have a sparse array.

If you have other values than one and zero in your array, more thought would be needed.