Stack 200 2-d numpy arrays of size (100*100) in a 3-d numpy array (200 * 100 * 100)

Question

I want to create 200 image frames of size (100*100) and stack them in a 3-D array with final size(200 * 100 *100) where x axis represent each frame ie, (1, 100, 100) should be the first frame.

I am not able to stack them in a 3-D array. The first loop does create a (2,100,100) array by stacking the first two frames but does not work after that and results in a (2,) array

import numpy as np
import random




def createCircle(width,height , rad ):
  w = random.randint(1, height)
  h = random.randint(1, height)
  center = [int(w), int(h)]
  radius = rad
  Y, X = np.ogrid[:height, :width]
  dist_from_center = np.sqrt((X - center[0])**2 + (Y-center[1])**2)
  mask = dist_from_center <= radius
  return mask


def addCircle(test_image):
  m = createCircle(width = 100, height = 100 , rad = 8 )
  masked_img = test_image.copy()
  masked_img[m] = 0
  return masked_img

img = np.zeros([100,100],dtype=np.uint8)
img.fill(20)
img_test = img  


def noise(image):
  row,col= image.shape
  mean = 0
  var = 0.1
  sigma = var**0.5
  gauss = np.random.normal(mean,sigma,(row,col))
  gauss = gauss.reshape(row,col)
  noisy = image + gauss #adding gauss noise
  s1 = np.sin(8)   #adding sin fill
  noisy += s1
  return noisy


#creates 1st frame
for i in range(4):
  im_first = addCircle(test_image=img_test)
  im_first = noise(im_first)



for i in range(200):
  for j in range(4):
    img_test = addCircle(test_image=img_test)
  im1 = noise(img_test)
  img_test = img
  im_first = np.array([im_first, im1])#stacks every new frame   (im1)#error in this    

I need a (200,100,100)

Answer 1

You can use numpy.stack for doing this. See my example

import numpy as np

a = np.array([[1,2],[3,4]])
b = np.array([[10,11],[12,13]])
c = np.stack((a,b))
print(c.shape) #prints (2,2,2)
print(c[0])
print(c[1])

output of 2nd print

[[1 2]
 [3 4]]

output of 3rd print

[[10 11]
 [12 13]]

Note that you have to feed tuple of 2D arrays into numpy.stack rather than individual 2D arrays

Answer 2

You can initialize an array and then fill it with the images. This tends to be more efficient than continuously stacking.

ims = np.zeros((200, 100, 100)) # initialize your array
for i in range(200):
  for j in range(4):
    img_test = addCircle(test_image=img_test)
  im1 = noise(img_test)
  ims[i, ...] = im1 # add the image to our initialized array

Answer 3

One approach would be to create a list of the 2-d matrices you want to stack and use np.stack(arrays, axis=0)

Example

arrays = [np.random.randn(100, 100) for _ in range(200)] #each matrix is of shape (100,100)
stacked_array = np.stack(arrays, axis=0) #stacked_array.shape is (200,100,100)