21.3. Working with Images

To read an image:

import cv2
image = cv2.imread('example.jpg')
cv2.imshow('Display Image', image)
cv2.waitKey(0)
cv2.destroyAllWindows()

• imread() – loads the image • imshow() – displays it in a window • waitKey(0) – waits for a key press • destroyAllWindows() – closes the window

This chunk covers the basic steps to read and display an image using OpenCV. First, you need to import the library with the statement import cv2. You then load an image from your computer by using the imread() function, providing the file name (like 'example.jpg') as an argument. The loaded image is stored in the variable image. To display the image in a window, use the imshow() function, which takes two arguments: the window title and the image variable. The program will continue until you press a key, controlled by waitKey(0), which waits indefinitely for a key press. Finally, to close the window, you call destroyAllWindows().

Imagine you have a photo album on your computer. To see a photo, you have to open the album, select a picture, and display it on your screen. In this analogy, cv2.imread() is like opening the photo album and choosing the picture, cv2.imshow() is like displaying the photo on your computer screen, and cv2.destroyAllWindows() is like closing the album after you're done looking at the pictures.

Images are stored as matrices of pixels:

• Grayscale images → 2D arrays (Height × Width) • Color images → 3D arrays (Height × Width × 3 channels - BGR)

In computer vision, images are actually just collections of data stored in a structured format called matrices. For grayscale images, which only contain shades of gray, the data is organized as a 2D array. Each element in this array represents the brightness of a pixel, where the dimensions are defined by the height and width of the image. Color images, on the other hand, are more complex and use a 3D array that includes three color channels: Blue, Green, and Red (often referred to as BGR). Each pixel's color is determined by its values in these three channels.

Think of an image like a large grid of colored squares, where each square represents a pixel. In a grayscale image, each square can only be a shade of gray, so it's like a two-dimensional chessboard where each square's brightness varies but only from black to white. For color images, imagine a 3D cube where each layer represents one color channel; every point in that cube contains information about the brightness of blue, green, and red for that pixel.

--