Import Libraries



In [1]:

    
import cv2



In [2]:

    
import matplotlib.pyplot as plt

Read Images



In [3]:

    
shooter = cv2.imread('demo1.jpg', cv2.IMREAD_GRAYSCALE)



In [4]:

    
nature = cv2.imread('nature.jpg', cv2.IMREAD_GRAYSCALE)



In [5]:

    
plt.figure(figsize=(20,15))









    Out[5]:





<matplotlib.figure.Figure at 0x7f12c336de10>



In [6]:

    
plt.subplot(1,2,1)









    Out[6]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f12c336de48>



In [7]:

    
plt.title('Shooter')









    Out[7]:





<matplotlib.text.Text at 0x7f12c3089320>



In [8]:

    
plt.imshow(shooter, cmap='gray')









    Out[8]:





<matplotlib.image.AxesImage at 0x7f12c30343c8>



In [9]:

    
plt.subplot(1,2,2)









    Out[9]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f12c3034320>



In [10]:

    
plt.title('Nature')









    Out[10]:





<matplotlib.text.Text at 0x7f12c3001be0>



In [11]:

    
plt.imshow(nature, cmap='gray')









    Out[11]:





<matplotlib.image.AxesImage at 0x7f12c2fb48d0>



In [12]:

    
plt.show()

Simple image Difference



In [13]:

    
img = shooter - nature



In [14]:

    
plt.imshow(img, cmap='gray')









    Out[14]:





<matplotlib.image.AxesImage at 0x7f12bee35f98>



In [15]:

    
plt.show()

Image looks absolutely unrecognizable here. It is due to the fact that, when d = (a-b<0), d = 0 because negative intensity value gets truncated to 0. Even, using abs() won't help, as a-b subtration will happen first and yeild 0 , and then abs(0) will happen, which has no meaning. This can be solved by using absdiff function of cv2.

Absolute Image Difference



In [16]:

    
img_abs = cv2.absdiff(shooter, nature)



In [17]:

    
plt.imshow(img_abs, cmap='gray')









    Out[17]:





<matplotlib.image.AxesImage at 0x7f12bedd0d68>



In [18]:

    
plt.show()

Finally,a little better image.