#segmentation

Question 1: Single choice

A robot must find a red ball in camera images taken under very different lighting, by thresholding pixel colours. Why is thresholding the hue in HSV usually more robust than thresholding the raw R, G and B values?

R, G and B all scale with the lighting; hue barely changes when only brightness does1
HSV stores each pixel with more bits than RGB, so small colour differences are easier to separate2
Hue is a linear combination of R, G and B, so thresholds on it are equivalent but simpler to choose3
Hue is undefined for dark pixels, so the shadows of the ball are removed automatically4

Timed test