Lossless compression of hyperspectral data obtained from Fourier-transform infrared imaging spectrometers

Hyperspectral data from a commercial Fourier-transform infrared imaging spectrometer is compressed using lossless Huffman coding. It is shown that, when using properly designed prediction schemes, data size can be significantly reduced even when using measurements from different instruments observing various scenes. Because acquisitions are often limited by disk write speed, such a compression also implies speed gains or less complex hardware in the acquisition and transmission chain.