Lossy data compression for imaging interferometer data using a wavelet transform-based image compression algorithm

Document Type


Journal/Book Title/Conference

Optical Spectroscopic Techniques and Instrumentation for Atmospheric and Space Research





Publication Date


First Page


Last Page



Data compression on future space-based imaging interferometers can be used to reduce high telemetry costs, provided the performance is acceptable. This paper investigates lossy data compression of imaging interferometer datacubes using a wavelet transform-based compression algorithm, the Set Partitioning in Hierarchical Trees (SPIHT) image compression algorithm. Compression is performed on individual frames of the interferogram datacubes. Simulated datacubes from the Geosynchronous Imaging Fourier Transform Spectrometer (GIFTS) are modified to produce new complex GIFTS datacubes used to perform the experiments. Separate programs are written for the encoder and decoder in C++. The encoder and decoder are simulated to the bit-level, meaning they simulate the exact bit streams that would be generated by hardware implementations. All compression ratios reported are based on the actual file size of the encoded data. The simulations indicate very high performance of the algorithm in the interferogram domain, with average errors of less than one least significant bit (LSB) for the GIFTS long-wave band and just over one LSB for the GIFTS short/mid-wave band at compression ratios as high as 13.7:1 and 15.4:1, respectively. At the same compression ratios, errors in the spectral radiance domain are comparable to the simulated instrument noise and RMS temperature profile retrieval errors of less than 1 K are achieved using a University of Wisconsin-Madison prototype retrieval algorithm.

This document is currently not available here.