How to add noise ( Gaussian /salt and pepper etc) to image in Python with OpenCV

There are several methods to add noise to images in Python using OpenCV. Common noise types include Gaussian noise and Salt and Pepper noise. The following sections detail how to add each of these noise types.

1. Gaussian Noise

Gaussian noise is a statistical noise type where the probability density function follows a normal distribution (Gaussian distribution). One common method involves generating a Gaussian random array of the same dimensions as the original image and adding it to the image.

python
import numpy as np
import cv2

# Read the original image
image = cv2.imread('path_to_image.jpg')

# Obtain the image dimensions
rows, cols, channels = image.shape

# Generate Gaussian noise
mean = 0  # Mean
var = 10  # Variance
sigma = var ** 0.5
gaussian = np.random.normal(mean, sigma, (rows, cols, channels))  # Create random array
gaussian = gaussian.reshape(rows, cols, channels)

# Add Gaussian noise to the original image
noisy_image = cv2.add(image, gaussian)

# Display images
cv2.imshow('Original', image)
cv2.imshow('Gaussian Noise', noisy_image)
cv2.waitKey(0)
cv2.destroyAllWindows()

2. Salt and Pepper Noise

Salt and Pepper noise is characterized by black and white speckles appearing on the image, simulating random pixel changes to black or white. This noise can be added by randomly selecting pixels and setting them to the maximum (255) or minimum (0) values.

python
import numpy as np
import cv2
import random

# Read the original image
image = cv2.imread('path_to_image.jpg')

# Obtain the image dimensions
rows, cols, channels = image.shape

# Salt and Pepper noise ratio
salt_pepper_ratio = 0.02
output = np.copy(image)

# Add salt noise
num_salt = np.ceil(salt_pepper_ratio * image.size * 0.5)
coords = [np.random.randint(0, i - 1, int(num_salt)) for i in image.shape]
output[coords[0], coords[1], :] = 255

# Add pepper noise
num_pepper = np.ceil(salt_pepper_ratio * image.size * 0.5)
coords = [np.random.randint(0, i - 1, int(num_pepper)) for i in image.shape]
output[coords[0], coords[1], :] = 0

# Display images
cv2.imshow('Original', image)
cv2.imshow('Salt and Pepper Noise', output)
cv2.waitKey(0)
cv2.destroyAllWindows()

In both methods, you can adjust parameters such as variance and noise ratio to achieve different levels of noise effects. These techniques are commonly used in image processing to test algorithm robustness or for enhancing training data.