How to edit a pixel using OpenCV

Recipe Objective: How to edit a pixel using OpenCV?

In this recipe, let us understand how OpenCV reads the images as pixels and how to manipulate those pixels

Comprehensive List of Computer Vision Project Ideas for Data Enthusiasts

Step 1: Import the library and read the image

The first step is to import the library and read the image using the cv2.imshow() function. The image that we are using for this recipe is as follows.

#Importing libraries
import cv2
import matplotlib.pyplot as plt

#Reading the image
image = cv2.imread('project.jpg')

Step 2: Understanding the structure of pixel array

If we look at the shape of the image variable, we can see that it is a multidimensional array of the shape (110, 335, 3)

print("The shape of the image is ",image.shape)

Output:

The shape of the image is  (110, 335, 3)

Here, 110 is the height, 335 is the width, and 3 is the number of channels (this is a color image, we have three channels in B, G, R format). Therefore, we have 110 X 335 = 36850 tuples of values each in (B, G, R). Each value in the tuple ranges from 0 to 255. Various combinations of these values form various colors. For instance,

• (255,255,255) corresponds to white color
• (0,0,0) corresponds to black
• (255,0,0) corresponds to red

Step 3: Edit a pixel

Let us try to change obtain the (B, G, R) value for one pixel and edit it to our convenience

(b,g,r) = image[20][50]

This gives us the (B, G, R) tuple value of the 50th pixel in the 20th row

print("The (B,G,R) value at 50th pixel of the 20th row is ",(b,g,r))

Output:

The (B,G,R) value at 50th pixel of the 20th row is  (81, 48, 32)

Now let us change pixel the value from (81,48,32) to (255,255,255)

image[20][50] = (255,255,255)

Step 4: Display the picture

Since we have changed only one pixel, it would be better to view the image using the plt.imshow() function of the matplotlib library. Matplotlib expects an RGB format while OpenCV uses BGR format for images. Hence let us convert the color scheme using the cv2.cvtColor() function, which takes the image and the color scheme as inputs.

As we want to convert the color scheme from BGR to RGB, we use cv2.COLOR_BGR2RGB code

plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))

Output:

If we look closer, we can see a tiny dot in white color above the logo, which was not previously present in the image