Compression technology for multi-channel telemetry data

The data obtained from the spectrometer are presented in the form of a three-dimensional structure depending on the type of spectrometer (Fourier or interference):

– the X and Y axes determine the spatial component;

– Z-axis – depends on the type of spectrometer and determines either the spectral component or a set of interferograms formed as a result of the aircraft movement and aligned in the spatial domain.

The 3D data is then decorrelated in both the spatial domain and the z-axis. This stage is necessary to reduce the dynamic range and prepare for the compression stage itself. To do this, a reference frame is subtracted along the Z axis, and after that  determination of an algorithm that will be used for decorrelation in the spatial domain is performed (for example, a prediction algorithm or 2D wavelet transform).

The decorrelated stream is passed to the entropy coding unit.

Various compression algorithms can be implemented within this general processing scheme.

Depending on the field of hyperspectral data application, the compressed stream is either stored in the computing system memory or transmitted over the radio channel to the flight control center.

Scientific and technical groundwork.

The prototype of the software system for compressing hyperspectral images based on contextual modeling for software and hardware systems for Earth remote sensing has been developed. The components of the compression system prototype are:

• hyperspectral cube reconstruction unit (generates PGM files, each of them contains one frame describing the spatial component);
• unit for decorrelation in spatial and spectral planes;
• entropy codec selection unit.