How to handle Data, and Images(13) OpenCV Filtering

Lesson Notes in .ipynb file

How to handle Data, and Images(13) OpenCV Filtering

Topics

• OpenCV Filtering
• Summary

OpenCV Filtering

• we can apply kernel to image to make the image more blurry
• when we make image more blurry it can reduce the noise of the image

Convolution

Basic Image: $\begin{bmatrix}5 &4 & 3 &2 & 5 \\ 4 & 6 & 5 & 4 & 3 \\ 7 & 5 & 2 & 5 & 9 \\ 8 & 5 & 6 & 1 & 9 \\ 5 & 7 & 7 & 8 & 6 \end{bmatrix} $

Kernel: $ \begin{bmatrix} 1 & 2 & 1 \\ 2 & 5 & 2 \\ 1 & 2 & 1\end{bmatrix}$

then for every element we multiply it by kernel.

For example for the output after convolution operation at [2][2] would be

$\begin{bmatrix} 6 & 5 & 4 \\ 5 & 2 & 5 \\ 5 & 6 & 1\end{bmatrix}$ * $\begin{bmatrix} 1 & 2 & 1 \\ 2 & 5 & 2 \\ 1 & 2 & 1\end{bmatrix}$ = $\begin{bmatrix} 6\times1 & 5\times2 & 4\times1 \\ 5\times2 & 2\times5 & 5\times2 \\ 5\times1 & 6\times2 & 1\times1 \end{bmatrix} $ = Total Sum After Multiplication / Total Sum of Kernel elements = $\frac{68}{17}$

Basic Kernel: where every element is the same * 1/sum of all elements

$\begin{bmatrix} 1 & 1 & 1 & 1 & 1\\ 1 & 1 & 1 & 1 & 1\\ 1 & 1 & 1 & 1 & 1\\ 1 & 1 & 1 & 1 & 1\\ 1 & 1 & 1 & 1 & 1 \end{bmatrix} * \frac{1}{25}$

Gaussian Kernel: matrix(highest in the middle and lower the value further from the middle) * 1/sum of all elements

$\begin{bmatrix} 1 & 4 & 7 & 4 & 1\\ 4 & 16 & 26 & 16 & 4\\ 7 & 26 & 41 & 26 & 7\\ 4 & 16 & 26 & 16 & 4\\ 1 & 4 & 7 & 4 & 1 \end{bmatrix} * \frac{1}{273}$

Basic Blurring by making our own basic Kernel

import cv2
import matplotlib.pyplot as plt
import numpy as np

image = cv2.imread('chanel.jpg')
plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
plt.show()

size = 25
kernel = np.ones((size, size), np.float32) / (size**2)

dst = cv2.filter2D(image, -1, kernel)
plt.imshow(cv2.cvtColor(dst, cv2.COLOR_BGR2RGB))
plt.show()

Output:

Base Image	Basic Blurring by making our own Basic Kernel

Basic Blurring

import cv2
import matplotlib.pyplot as plt

image = cv2.imread('chanel.jpg')
plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
plt.show()

dst = cv2.blur(image, (25, 25))
plt.imshow(cv2.cvtColor(dst, cv2.COLOR_BGR2RGB))
plt.show()

Output:

Base Image	Basic Blurring

Gaussian Blur

import cv2
import matplotlib.pyplot as plt

image = cv2.imread('chanel.jpg')
plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
plt.show()

# kernel_size must be odd number
dst = cv2.GaussianBlur(image, (25, 25), 0)
plt.imshow(cv2.cvtColor(dst, cv2.COLOR_BGR2RGB))
plt.show()

Output:

Base Image	Gaussian Blurring

Summary

We can blur an image using Convolution operation towards the image

There are two types of Kernel we can apply for Image Blurring

• Basic Kernel: where every element is the same * 1/sum of all elements
  • cv2.blur(image, (x, x)), where $(x,x)$ is a size of Kernel
• Gaussian Kernel: matrix(highest in the middle and lower the value further from the middle) * 1/sum of all elements
  • cv2.GaussianBlur(image, (x, x), 0), where $(x,x)$ is a size of Kernel, and $x$ must be odd number

The bigger the size of Kernel more blurry the image gets