Which version of Python and OpenCV should I use for resizing an image?

You can use the Python version 3.6.0 and the OpenCV version 3.2.0.

Should I know the basics of Python programming before downloading the approved versions?

It would be best if you already know the basics of Python programming.

Are there any important operations of image processing in OpenCV?

Yes, you should check the Core Operations in OpenCV for image processing.

How to Resize an Image in Python (+ Examples)

In this article, we will cover the basics of image manipulation in OpenCV and how to resize an image in Python, its cropping, and rotating techniques.

What will we cover in this article?

We will use Python version 3.6.0, OpenCV version 3.2.0. It is an assumption that you have Python installed on your machine and already know the basics of Python programming. Please read the article if you need a tutorial on how to install OpenCV for Python.

Before we go any further, let’s remember about Core Operations in OpenCV for image processing. This operation consists of reading the image, displaying the image, and saving the image.

Dopinger · How To Resize An Image In Python

Resize Images in Bulk

Image processing and resizing have an essential value in SEO, Page Speed, and Bandwidth optimization. For these reasons, every year, there are new developments in coding. In this section, we’ll share with Phyton how to resize multiple images in bulk. With Phyton, you can read the size and compression of ideas in bulk in the rest of our article.

For these operations, we need to use the Pillow Library first. If you have not downloaded Pillow to your system before, you can use the following code:

pip install pillow

First, we need to import Pillow Libraries:

from PIL import Image
import os
import PIL
import glob

To use shorter file names in our system, we should prefer OS and Glob modules. It is used to change, save and open images with PIL to PIL image or different optimization options.

image = Image.open("pic.jpeg")
print(image.size)

OUTPUT>>>

(2000, 1700)

resized_image = image.resize((600,600))
print(resized_image.size)

OUTPUT>>>

(600, 600)

Reading Images

To read images in OpenCV, use a function cv2.imread()where the first parameter is the image file name complete with its extension. As an example :

import cv2
img = cv2.imread('pic.jpg')

Cv2 function. Whiskey ( 0 ) is to keep the window displaying the image. While the cv2 function. destroyAllWindows ( ) is to close other windows that are currently open.

Displaying an Image

Displaying an image in OpenCV using a function cv2.imshow()where the first parameter is the window name to display the image and the second parameter is the image itself.

import cv2

img = cv2.imread('pic.jpg')
cv2.imshow('Displaying Images', img)

Writing / Saving Images

To write / save images in OpenCV using a function cv2.imwrite()where the first parameter is the name of the new file that we will save and the second parameter is the source of the image itself.

import cv2

img = cv2.imread('pic.jpg')
cv2.imwrite('img1.jpg', img)

For details on OpenCV Core Image Operations, please read the OpenCV documentation.

Now we can go back to the original topic of basic image manipulation in OpenCV and Python. As explained earlier in this article, we will learn how to apply resizing, cropping, and rotating techniques to images.
Let’s first try reading our image source and displaying it with the functions previously described.

import cv2

img = cv2.imread("pic.jpg")
cv2.imshow("Foo Window", img)
cv2.waitKey(0)
cv2.destroyAllWindows()

Henceforth, we will use the image above in this paper. Run the print command ( img . Shape ) to display the dimensions of our source image. The command will output (680, 850, 2) where 680 is the width, and 850 is the height in pixel size, while 2 is the image channel (RGB), or it means that the image has 680 rows and 850 columns.

Please note that if we read the image in grayscale form, the output will only produce rows and columns. So the img command . shape can also be applied to see if the image is grayscale or color image.

Image Resizing OpenCV

Earlier we got the width of our image with the img function . shape that is 850 pixels. Let’s resize the image to be 2 times smaller.

import cv2

img = cv2.imread("pic.jpg")
h, w = img.shape[:2]

#Specifying Image Size and Resizing

new_h, new_w = int(h / 2), int(w / 2)
resizeImg = cv2.resize(img, (new_w, new_h))

#Displaying Images

cv2.imshow('Original', img)
cv2.imshow('Resizing', resizeImg)
cv2.waitKey(0)
cv2.destroyAllWindows()

OpenCV Image Cropping

Keep in mind that every image we read with the cv2.imshow () function returns data in the form of an array.

Cropping application to OpenCV is very easy; we need to determine where the coordinates of the image to be cropped. First, we determine the initial x coordinate and final x, then determine the initial y coordinate and end y coordinates of the image that has been said to be read earlier.

croppedImg = img[420:540, 750:860]
cv2.imshow('Resizing', croppedImg)
cv2.waitKey(0)

From the command above, the crop results from our initial image will appear following the coordinates we specified earlier.

Image Rotating OpenCV

Changing the rotation isn’t that difficult either. First we have to determine the center point of rotation which we can determine from the width and height of the image, then determine the degree of rotation of the image and the dimensions of the image output.

img = cv2.imread('pic.jpg')
h,w = img.shape[:2]
center = (w/2,h/2)
rotate = cv2.getRotationMatrix2D(center,170,1)

Next is to apply the rotation settings that we have defined on the image we read earlier and display the image.

rotatingImg = cv2.warpAffine(img,rotate,(w,h))
cv2.imshow('Rotating', rotatingImg)
cv2.waitKey(0)

Quick Recap on Comands

Here’s a list that will help you refresh you memory.

Reading Images

How to Resize an Image in Python in Short

These are the basics of Image Manipulation with OpenCV and the ways you can resize an image in Python. You can easily make arrangements with the image sizes in Python.