Share Email Print
cover

Proceedings Paper

Wavelet-based coding of ultraspectral sounder data
Author(s): Fernando Garcia-Vilchez; Joan Serra-Sagrista; Francesc Auli-Llinas
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

In this paper we provide a study concerning the suitability of well-known image coding techniques originally devised for lossy compression of still natural images when applied to lossless compression of ultraspectral sounder data. We present here the experimental results of six wavelet-based widespread coding techniques, namely EZW, IC, SPIHT, JPEG2000, SPECK and CCSDS-IDC. Since the considered techniques are 2-dimensional (2D) in nature but the ultraspectral data are 3D, a pre-processing stage is applied to convert the two spatial dimensions into a single spatial dimension. All the wavelet-based techniques are competitive when compared either to the benchmark prediction-based methods for lossless compression, CALIC and JPEG-LS, or to two common compression utilities, GZIP and BZIP2. EZW, SPIHT, SPECK and CCSDS-IDC provide a very similar performance, while IC and JPEG2000 improve the compression factor when compared to the other wavelet-based methods. Nevertheless, they are not competitive when compared to a fast precomputed vector quantizer. The benefits of applying a pre-processing stage, the Bias Adjusted Reordering, prior to the coding process in order to further exploit the spectral and/or spatial correlation when 2D techniques are employed, are also presented.

Paper Details

Date Published: 19 August 2005
PDF: 10 pages
Proc. SPIE 5889, Satellite Data Compression, Communications, and Archiving, 58890K (19 August 2005); doi: 10.1117/12.616950
Show Author Affiliations
Fernando Garcia-Vilchez, Univ. Autonoma de Barcelona (Spain)
Joan Serra-Sagrista, Univ. Autonoma de Barcelona (Spain)
Francesc Auli-Llinas, Univ. Autonoma de Barcelona (Spain)


Published in SPIE Proceedings Vol. 5889:
Satellite Data Compression, Communications, and Archiving
Bormin Huang; Roger W. Heymann; Charles C. Wang, Editor(s)

© SPIE. Terms of Use
Back to Top