🖼️ Image Representation: Grayscale, RGB, etc.

When we talk about image representation, we mean how an image is stored inside a computer—in terms of data structures, colour channels and pixel values. This representation directly affects how we process, analyse and enhance the image later.

---

1. 🌗 Grayscale Images

A grayscale image is represented as a 2D matrix of light intensity values:

• 0 = black
• 255 = white
• Intermediate values = shades of grey

Example (3×3 matrix):

[   0  127  255 ]
[  64  128  192 ]
[  30   60   90 ]

💡 Use case: Faster processing when colour information isn’t required (e.g. edge detection).

2. 🌈 RGB Images (Colour Images)

A colour image in the RGB model uses three channels:

• R = Red
• G = Green
• B = Blue

Each pixel is represented by three values (one per channel) and stored as a 3D array:

Height × Width × 3

Example (single pixel):

R: 120  
G:  60  
B:  30  

→ This produces a brown-orange colour.

🖥️ Visual Layout

Conceptually, an RGB image is like stacking three grayscale images (one per channel) on top of each other.

🐍 NumPy / OpenCV Shape

In Python/OpenCV, you can inspect the shape of loaded images:

import cv2

img_gray = cv2.imread("bird.jpg", cv2.IMREAD_GRAYSCALE)
img_rgb  = cv2.imread("bird.jpg", cv2.IMREAD_COLOR)

print(img_gray.shape)  # e.g. (300, 400)
print(img_rgb.shape)   # e.g. (300, 400, 3)

3. 🧰 Other Formats (Preview Only)

Covered in the next topic on colour spaces.

Format	Description	Use Case
HSV	Hue, Saturation, Value	Colour filtering, tracking
YCrCb	Luma (Y) and chroma (Cr, Cb) channels	Video compression
Lab	Lightness and opponent-colour axes	Colour difference metrics

---

🔄 Summary

Image Type	Dimensions	Values	Common Use
Grayscale	H × W	0–255	Simpler, faster processing
RGB	H × W × 3	3 channels	Full-colour analysis
HSV/YCrCb	H × W × 3	Converted	Colour filtering, compression