Generate average image using Python and PIL (Python Image Library)

This page shows how to generate an average image of the image arrays using python and PIL (python image library) module.
It is easy to do by converting the image to the numpy.array.

In [1]:

%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np
from PIL import Image

Generate the original images

In [2]:

for i in range(1, 21):
    col = int(255 * i / 20)
    img = Image.new('RGB', (120, 80), color=(col, col, col))
    img.save('%02d.bmp' % i, format='bmp')

Load the original images saved just above

In [3]:

ims = []
for i in range(1, 21):
    ims.append(Image.open('%02d.bmp' % i, mode='r'))

Check the original images

In [4]:

ims[0]

Out[4]:

In [5]:

ims[10]

Out[5]:

In [6]:

ims[-1]

Out[6]:

Convert the loaded images to numpy ndarray

In [7]:

ims = np.array([np.array(im) for im in ims])

Check the shape of the ims array

In [8]:

ims.shape

Out[8]:

(20, 80, 120, 3)

Generate the averaged array

In [9]:

imave = np.average(ims,axis=0)

Check the shape of the ims array

In [10]:

imave.shape

Out[10]:

(80, 120, 3)

Convert the averated image (in float) to PIL image (uint8)

In [11]:

result = Image.fromarray(imave.astype('uint8'))

Save the averated uint8 image

In [12]:

result.save('result.bmp')