In the past few years there has been a significant volume of research work carried out in the field of multispectral image
acquisition. The focus of most of these has been to facilitate a type of multispectral image acquisition systems that
usually requires multiple subsequent shots (e.g. systems based on filter wheels, liquid crystal tunable filters, or active
lighting). Recently, an alternative approach for one-shot multispectral image acquisition has been proposed; based on an
extension of the color filter array (CFA) standard to produce more than three channels. We can thus introduce the concept
of multispectral color filter array (MCFA). But this field has not been much explored, particularly little focus has been
given in developing systems which focuses on the reconstruction of scene spectral reflectance.
In this paper, we have explored how the spatial arrangement of multispectral color filter array affects the acquisition
accuracy with the construction of MCFAs of different sizes. We have simulated acquisitions of several spectral scenes
using different number of filters/channels, and compared the results with those obtained by the conventional regular MCFA
arrangement, evaluating the precision of the reconstructed scene spectral reflectance in terms of spectral RMS error, and
colorimetric ▵E*ab color differences. It has been found that the precision and the the quality of the reconstructed images
are significantly influenced by the spatial arrangement of the MCFA and the effect will be more and more prominent with
the increase in the number of channels. We believe that MCFA-based systems can be a viable alternative for affordable
acquisition of multispectral color images, in particular for applications where spatial resolution can be traded off for spectral
resolution. We have shown that the spatial arrangement of the array is an important design issue.