Working with Images

In this section, we’ll dive deep into how OpenCV handles image data, color spaces, and pixel manipulation. Mastering these concepts is key to tackling any advanced computer vision project.

🎨 Color Spaces in OpenCV

OpenCV loads images in BGR (Blue-Green-Red) by default. However, for different applications, you often need to convert images to other color spaces like:

RGB – for accurate color display
Grayscale – for light intensity only
HSV – separates color from brightness
LAB – simulates human color perception

🔁 `cv2.cvtColor()`

This function converts an image from one color space to another.

Syntax:

cv2.cvtColor(src, code)

Parameters:

src: Input image
code: Conversion code (e.g. cv2.COLOR_BGR2GRAY)

Returns: New image in the target color space.

Convert BGR to Grayscale

import cv2
print("Reading image for grayscale conversion...")
img = cv2.imread('sample.jpg')
if img is None:
    print("Failed to load image.")
else:
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    print("Image converted to grayscale.")
    cv2.imwrite('gray.jpg', gray)

📷 Original Image

🖼️ Grayscale Output

Convert BGR to HSV

import cv2
print("Reading image for HSV conversion...")
img = cv2.imread('sample.jpg')
if img is None:
    print("Failed to load image.")
else:
    hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
    print("Image converted to HSV.")
    cv2.imwrite('hsv.jpg', hsv)

📷 Original Image

🖼️ HSV Output

HSV separates color (hue) from intensity and saturation, making it ideal for color-based segmentation and filtering.

🧮 Accessing and Modifying Pixels

Every image in OpenCV is stored as a NumPy array. You can access pixels like this:

Read a Pixel Value

import cv2
img = cv2.imread('sample.jpg')
b, g, r = img[100, 50]  # row=100, col=50
print("Blue:", b, "Green:", g, "Red:", r)
#Output: Blue: 14 Green: 0 Red: 1

Modify a Pixel Value

import cv2
img = cv2.imread('sample.jpg')
img[100, 50] = [0, 0, 255]  # Set to red (BGR)
cv2.imwrite('modified_pixel.jpg', img)

📷 Modified Image Output

Now, it seems like the image hasn’t changed. But if you look closely (by zooming in alot!), you can see that the pixel at (100, 50) is now red.

🔍 Image Properties

OpenCV images have the following attributes:

Property	Description	Example
`shape`	Tuple of (height, width, channels)	`(512, 512, 3)`
`size`	Total number of elements	`786432`
`dtype`	Data type of image elements	`uint8`

import cv2
img = cv2.imread('sample.jpg')
print("Shape:", img.shape)
print("Size:", img.size)
print("Data type:", img.dtype)

Output:
Shape: (3306, 4959, 3)
Size: 49183362
Data type: uint8

Most images are stored as 8-bit unsigned integers (uint8). This means each color channel ranges from 0–255.

📝 Practice Questions

❓ Question 1

Read an image and convert it to LAB color space. Save the result.

Show Code

import cv2
img = cv2.imread('sample.jpg')
lab = cv2.cvtColor(img, cv2.COLOR_BGR2LAB)
cv2.imwrite('lab.jpg', lab)

📷 LAB Image Output

❓ Question 2

Access pixel at (150, 80) and print its BGR values.