Selection of Training Sets for the Characterisation of multispectral Imaging Systems

Deriving the actual multispectral data from the output of the acquisition system is a key problem in the field of multispectral imaging. Solving it requires a correlation method and the training set (if any) on which the method relies. In this paper we propose two novel approaches in selecting a training set to be used for the characterisation of a multispectral acquisition system. In both cases the selected training sets will have low numerosity and broad applicability. We also test both approaches on the data obtained from a real acquisition, comparing the reconstructed reflectances with the measurements obtained using a spectrophotometre.