Operations with Images in OpenCV

To use OpenCV in Python, you need to install the openCV module using conda command.

conda install opencv

Now you can import OpenCV into your Python code.

import cv2 as cv

Input/Output

To load an image from a file:

# Load 'test.jpg' image from your current directory
img = cv.imread("test.jpg")
# Load 'test.jpg' image from 'pictures' directory
img = cv.imread("pictures/test.jpg") 

# Load image from 'filename' variable
filename = "test.jpg"
img = cv.imread(filename)

If you read a jpg file, a 3 channel image is created by default. If you need a grayscale image, use:

filename = "test.jpg"
# read test.jpg but it read our image as grayscale
img = cv.imread(filename, cv.IMREAD_GRAYSCALE)

# print shape of image
print(img.shape)

To save/write your image from OpenCV:

# write 'img' to your path
cv.imwrite('test2.jpg', img)

To show your image:

# show 'img' with title name 'test show image'
cv.imshow('test show image',img)
# show image for 5 seconds (5000 ms) before it automatically close it
cv.waitKey(5000)
# or you can use this line, it will wait until you pressed any key
cv.waitKey(0)

cv.destroyAllWindows()

Color Spaces in OpenCV

Color spaces are a way to represent the color channels present in the image that gives the image that particular hue. BGR color space: OpenCV’s default color space is RGB. However, it actually stores color in the BGR format.

To visualize the different color channels of RGB image

import cv2 as cv

# you can use any image on your computer
img = cv.imread('cat.jpg')
# seperate each channel of RGB image
B, G, R = cv.split(img)

# show each channel
cv.imshow("original", img)
cv.waitKey(0)
 
cv.imshow("blue", B)
cv.waitKey(0)
 
cv.imshow("Green", G)
cv.waitKey(0)
 
cv.imshow("red", R)
cv.waitKey(0)
 
cv.destroyAllWindows()

Arithmetic Operations on Images

We can applied arithmetic operations like Addition, Subtraction, and Bitwise Operation to our input images.

Addition

cv.add(img1, img2)

Subtraction

cv.subtract(img1, img2)

Bitwise

Note: In bitwise operation, both images must have the same dimensions.

Parameters for AND, OR, XOR source1 : First Input Image array(Single-channel, 8-bit or floating-point) source2 : Second Input Image array(Single-channel, 8-bit or floating-point) dest : Output array (Similar to the dimensions and type of Input image array) mask: Operation mask, Input / output 8-bit single-channel mask

dest_and =  cv.bitwise_and(img1, img2, mask=None)
dest_or = cv.bitwise_or(img2, img1, mask = None)
dest_xor = cv.bitwise_xor(img1, img2, mask = None)

Parameters for NOT source : Input Image array(Single-channel, 8-bit or floating-point) dest : Output array (Similar to the dimensions and type of Input image array) mask : Operation mask, Input / output 8-bit single-channel mask

dest_not1 = cv.bitwise_not(img1, mask = None)
dest_not2 = cv.bitwise_not(img2, mask = None)

##Image Processing in Python

Scaling an Image

import cv2 as cv
import numpy as np

filename = 'cat.jpg'
img = cv.imread(filename)

# Get number of pixel horizontally and vertically
(height, width) = img.shape[:2]

res = cv.resize(img, (int(width / 2), int(height / 2)), interpolation = cv.INTER_CUBIC)

cv.imwrite("result.jpg", res)

Now you can check the dimensions of original and result image.

Rotating an Image

import cv2 as cv
import numpy as np

filename = 'cat.jpg'
img = cv.imread(filename)

# Shape of image in terms of pixels.
(rows, cols) = img.shape[:2]

# getRotationMatrix2D creates a matrix needed for transformation.
# We want matrix for rotation w.r.t center to 45 degree without scaling.
M = cv.getRotationMatrix2D((cols / 2, rows / 2), 45, 1)
res = cv.warpAffine(img, M, (cols, rows))

cv.imshow("rotated!", res)
cv.waitKey(0)

Acknowledgement : The content of this document has been adapted from the original OpenCV Python Tutorial from GeeksforGeeks