Rgb And Hsv Colour Spaces

RBG and HSV colour spaces

[2021-09-27]

We, humans sense light and colours in RGB, i,e, using Red, Green, and Blue light sensors of our eyes. But we consciously understand light better in HSV, i.e. Hue, Saturation, and Value.

For example, when I tell you that a colour pixel is [100, 160, 35] to represent the red, green, and blue 8-bit colour channels, what colour do you picture? Contrast that with [0.25, 0.64, 0.38] to represent the hue, saturation, and value of a pixel. We know that for hue, i.e. the colour wheel goes across the 3 primary colours where red is at position 0, green at a third of the way (i.e. ~33% or 120°), and blue at two-thirds of the way (i.e. ~67% or 240°). The colours transitions from red, orange, yellow, yellow-green, green, cyan, blue, violet, magenta, then back to red. Now since we have 25% hue value or 90° along the colour wheel, then we can say that the pixel is yellow-greenish.

Python script to generate the RGB and HSV colour representations:

########################
### Import libraries ###
########################
import numpy as np
from skimage import color
import matplotlib; matplotlib.use('TkAgg')
from matplotlib import pyplot as plt

###########
### RGB ###
###########
### range of channel values
n_x_cols = 256
n_y_rows = 256
n_z_planes = 5
### vector of channel values
vec_R = np.uint8(np.arange(0, 256, 1))
vec_G = np.uint8(np.arange(0, 256, 1))
vec_B = np.uint8(np.arange(0, 256, int(256/n_z_planes)))
### initialise colour image
mat_RG = np.uint8(np.zeros((n_x_cols, n_y_rows, 3)))
### set values along the red channel
for x in range(n_x_cols):
  mat_RG[:,x,0] = vec_R[x] ### red
### set values along the green channel
for y in range(n_y_rows):
  mat_RG[y,:,1] = vec_G[y] ### green
### set values along the blue channels and save the image
for z in range(len(vec_B)):
  mat_RG[:,:,2] = vec_B[z] ### blue
  ### plot
  plt.imshow(mat_RG, origin='lower')
  plt.ylabel("Green")
  plt.xlabel("Red")
  # plt.show(block=False)
  plt.savefig("RGB-"+str(vec_B[z])+".svg")

###########
### HSV ###
###########
### range of channel values
n_x_cols = 100
n_y_rows = 100
n_z_planes = 10
### vector of channel values
vec_H = np.arange(0, 100, 1)/100
vec_S = np.arange(0, 100, 1)/100
vec_V = np.arange(0, 100, int(100/n_z_planes))/100
vec_V = np.append(vec_V, 1.0)
### initialise colour image
mat_HS = np.zeros((n_x_cols, n_y_rows, 3))
### set values along the hue channel
for x in range(n_x_cols):
  mat_HS[:,x,0] = vec_H[x] ### hue
### set values along saturation channel
for y in range(n_y_rows):
  mat_HS[y,:,1] = vec_S[y] ### saturation
### set values along the values channel and save the image
for z in range(len(vec_V)):
  mat_HS[:,:,2] = vec_V[z] ### value
  ### plot
  plt.imshow(color.hsv2rgb(mat_HS), origin='lower')
  plt.xlabel("Hue")
  plt.ylabel("Saturation")
  # plt.show(block=False)
  plt.savefig("HSV-"+str(vec_V[z])+".svg")