Computational Colour Science Using MATLAB, 2nd Edition

Chapter 10

Characterisation of Cameras

10.1 Introduction

Characterisation is only worthwhile assuming that the device has already been calibrated. In the case of a camera system, this would normally mean that the imaging system, in a wider sense, is controlled. For example, the system could be configured so that the subject is a fixed distance from the camera lens and is illuminated uniformly. The camera itself should have all settings (focus, exposure time, white balance) optimised and fixed. The purpose of calibration is so that the camera is capable of capturing consistent RGB values over time. White balance deserves special attention and is a process of adjusting the colour balance in an image to remove unwanted colour castes. Most digital SLR cameras have three settings for white balance; automatic, preset and custom. Automatic white balance is where the camera on-board software uses a ‘best-guess’ algorithm to estimate the colour of the illumination and correct for it. Preset white balance is where the camera uses a fixed (preset) estimation of the illumination (such as daylight or tungsten). Custom white balance is where the user can take an image of, say, a grey reference card under lighting that will subsequently be used for images. It is very important that automatic white balance (which is the default for many digital cameras) is not used. Custom white balance is obviously appropriate if images are being captured in, say, a viewing cabinet under stable illumination. ...