Advanced remote-sensing instruments in recent years often employ multiple wavelengths for maximal science return. Many of them comprise hundreds to over a thousand spectral bands. These instruments multiply the data volume and stress the communication bandwidth. With previously deployed space lossless data compression standards only applicable to individual bands, a need exists to further improve performance by using spectral information.
The Lossless Multispectral & Hyperspectral Image Compression standard published by the Consultative Committee on Space Data Systems (CCSDS) in 2012 is a direct response to the need. The specified algorithm exploits the correlation among adjacent spectral bands and utilizes an adaptive filtering technique to improve compression performance. The additional gain can be from 10 to over 30%, depending on the data, while preserving the full fidelity of the data.
A C-code is produced to implement this CCSDS standard. It allows users to choose the number of bands to use in the prediction stage as well as one of the two specified entropy coders and their associated parameter values. The maximal specified input dynamic range is 16-bit and the coded bit sequence is fully compliant with the CCSDS specifications.