Chapter 14. Colour.
So far in this book all the discussions on techniques for computer vision have assumed that the
sensing system receiving the incoming energy is achromatic or monochromatic. It is now time to
take a look at the extraction of the extra information that colour can provide in an image. Without
such information a variety of tasks would become very difficult to perform; for example, determining
the colour of, say, a light that might indicate danger solely depending on its colour. It would also be
useful in a control structure that had a priori information of a scene. For example, if one is looking
for a particular coloured object, and one knows that the colour of that object contrasts with the
colours of the rest of the scene, then the search can become more constrained, faster and more
accurate, regardless of the luminance distribution.
Over the past few years there have been several attempts at providing different representations of
colour information - some driven by attempts to interpret biological colour processing and some
based more on a mathematical appreciation of the physics of colour. At the same time visual