I have an image defined as:
imagen = scipy.misc.imread("C:\\Users\\Reymi\\Downloads\\imagen.png")
It has a size of (1929,1280) I need to add 2 columns and rows with zeros to each side of the array so I can use a kernel without worrying about it going outside of the array.
How? Thank you
Use the pad function:
im_new = scipy.pad(array=imagen, pad_width=[2, 2], mode='constant', constant_values=0)
The above function makes the array bigger by 2 in all directions. If you wanted to replace the outer two pixels with 0's without making the picture larger I don't think there is a built-in function, but I would do something like:
pad = 2
im_new = scipy.zeros_like(imagen)
im_new[pad:-pad, pad:-pad] = imagen[pad:-pad, pad:-pad]
This creates an image of 0's, then fills in the middle with the middle part of imagen.
Option - 1 - Using OpenCV
imagenBorder = cv2.copyMakeBorder(imagen,2,2,2,2,cv2.BORDER_REFLECT)
Option - 2 - Manually padding to numpy array
import numpy as np
borderWidth = 2
i = 0
while i < borderWidth:
imagen = np.insert(imagen, 0, 0, axis=1)
imagen = np.insert(imagen, imagen.shape[1], 0, axis=1)
imagen = np.insert(imagen, 0, 0, axis=0)
imagen = np.insert(imagen, imagen.shape[0], 0, axis=0)
i = i+1
