Multispectral loss-less compression using approximation methods

The large size of multispectral data files is currently a major issue in multispectral imaging. The transmission of multispectral data over networks, as well as the storage of large archives, are strongly limited, so that a clear need for good compression methods arises. In this paper, we explore the possibility of loss-less compression for multispectral data through a number of approximation methods that operate on the spectral domain. To evaluate the performance of these methods, we apply them to a representative spectra database, and consider the corresponding decrease in information entropy as well as the classical file size ratio.