Image Representation in OpenCV | Pixels, RGB, Grayscale, HSV Explained

Lesson 1.3: Image Representation – Pixels, RGB, Grayscale, HSV

In computer vision, an image is represented as a matrix of pixels. Each pixel contains intensity values that define its brightness and color. Understanding pixels, RGB, Grayscale, and HSV is the foundation of image processing in OpenCV.

🟨 What are Pixels?

A pixel (picture element) is the smallest unit of an image. Images are stored as a grid of pixels, where each pixel has values that represent either intensity (black & white) or color (RGB/HSV).

🌈 RGB (Red, Green, Blue)

The RGB model represents images using three channels: Red, Green, and Blue. Each pixel has a combination of values (0–255) for each channel. For example: (255, 0, 0) is pure red, (0, 255, 0) is pure green, and (0, 0, 255) is pure blue.

import cv2
import matplotlib.pyplot as plt

img = cv2.imread("image.jpg")   # Reads in BGR format
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

plt.imshow(img_rgb)
plt.title("RGB Image")
plt.show()

⚫ Grayscale

A Grayscale image uses only one channel (0–255 intensity values). It removes color information and keeps brightness only, which is useful for tasks like edge detection and thresholding.

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
cv2.imshow("Grayscale", gray)
cv2.waitKey(0)
cv2.destroyAllWindows()

🎨 HSV (Hue, Saturation, Value)

The HSV model separates image color into three components:
🔹 Hue – type of color (red, green, blue, etc.) 🔹 Saturation – intensity of the color 🔹 Value – brightness of the pixel HSV is often used in color detection and tracking.

hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
cv2.imshow("HSV Image", hsv)
cv2.waitKey(0)
cv2.destroyAllWindows()

✅ To summarize:
• Pixels are the smallest building blocks of an image.
• RGB uses Red, Green, and Blue channels.
• Grayscale removes color and keeps intensity only.
• HSV separates colors into hue, saturation, and brightness.

In the next lesson, we will explore how to read, write, and display images in OpenCV.