Spectral-based camera characterisation

Spectral-based camera characterisation can be further divided into two methods  

The second method is to predict the surface reflectance. The surface reflectance function r is approximated by a linear model  

A number of researchers used this methodology (Cohen, 1964, 1988 D Zmura and Lennie, 1986 Marimont and Wandell, 1992 Westland and Thomson, 2000) to predict spectral reflectance. The results are summarised as follows  

In general, too many basis functions will produce spectral reflectance functions too ragged to be realistic. Methods for overcoming this phenomenon have been only partially successful. A model with three basis functions will give the reflectance function with some smoothness properly, but the reflectance function cannot be assured within the desired range (less than one and greater than zero). 

Alternatively, there are other methods which can also predict spectral reflectance functions such as those based on a smoothness parameter (Li and Luo, 2001a) and a colour inconstancy index (Luo et al, 1999). The DigiEye system appUes these two measures to predict the spectral reflectance functions. The smoothness approach is to minimise the difference between the values in 

---

Alternatively, Spectral Based Camera Characterisation can be used. This can be divided into two measurement of the spectral sensitivities of the camera and recovery of the spectral reflectances. Both will be introduced later. 

---