Crop image OpenCV

Open Source Computer Vision (OpenCV) is an open-sourceA software or a library whose code is made available to the public. computer vision and machine learning software library which provides a wide range of functions and algorithms to process and analyze images and videos.

Note: You can learn more about OpenCV.

Image matrix

An image is typically represented as a two-dimensional matrix, where each element corresponds to a pixel value within the image. These pixel values store information regarding the color intensity of each pixel.

In the case of grayscale images, each pixel value ranges from 0 to 2550 \textnormal{ to } 255. For example, a greyscaled image of size 4×34 \times 3 pixels will be:

[[135, 150, 120, 100],

 [50, 80, 200, 90],

 [200, 180, 220, 75]]
Image matrix of a grayscaled image

However, for colored images, each pixel value is represented by an RGB (Red, Green, Blue) matrix in the form [R, G, B]. The R, G, and B values individually range from 0 to 2550 \textnormal{ to } 255, indicating the intensity of the respective color channels. For example, a colored image of dimensions 2×2×32 \times 2 \times 3 pixels might have an image matrix like the following:

[[[255, 0, 146], [0, 255, 45]],

 [[0, 0, 255], [255, 255, 69]]]
Image matrix of a colored image

Cropping an image

Cropping an image means selecting and extracting a specific rectangular region from the original image. Python library OpenCV can do this job efficiently as follows:

import cv2

# Load the original image
original_image = cv2.imread('image.jpg')

# Define the dimensions of the output image
output_width = 400
output_height = 300

# Calculate the starting coordinates for cropping
start_x = (original_image.shape[1] - output_width) // 2
start_y = (original_image.shape[0] - output_height) // 2

# Perform the cropping
cropped_image = original_image[start_y : start_y + output_height, start_x : start_x + output_width]

# Display or save the cropped image
cv2.imshow('Original Image',original_image)
cv2.imshow('Cropped Image', cropped_image)
cv2.waitKey(0)
cv2.destroyAllWindows()
Python code to crop an image

Note: You can drag the image windows to have a better comparison.

Code explanation

  • Line 1: Imports cv2 library.

  • Line 4: Loads the image (named image.jpg) that's supposed to be cropped and stores it in original_image. This image has the dimensions of 700×700700 \times 700 pixels.

  • Lines 7–8: output_width and output_height store the dimensions of the resultant cropped image.

  • Lines 11–12: start_x stores the starting x-coordinate of the point from where the image will be cropped. start_y does the same thing for y-coordinate.

  • Line 15: Slices the original image matrix (original_image) and stores it in cropped_image. Now, the cropped_image variable contains the image matrix of the cropped image.

  • Line 18: Displays the uncropped image.

  • Line 19: Displays the cropped image.

  • Line 20: waitKey() method waits for the user to press any key to close all the windows.

  • Line 21: Closes all the opened windows at the end of the program.

  • and storing it in original_image variable.

Learn more about it:

Free Resources

Copyright ©2026 Educative, Inc. All rights reserved