Paper
24 September 2013 Improving multispectral satellite image compression using onboard subpixel registration
Mathieu Albinet, Roberto Camarero, Maxime Isnard, Christophe Poulet, Jokin Perret
Author Affiliations +
Abstract
Future CNES earth observation missions will have to deal with an ever increasing telemetry data rate due to improvements in resolution and addition of spectral bands. Current CNES image compressors implement a discrete wavelet transform (DWT) followed by a bit plane encoding (BPE) but only on a mono spectral basis and do not profit from the multispectral redundancy of the observed scenes. Recent CNES studies have proven a substantial gain on the achievable compression ratio, +20% to +40% on selected scenarios, by implementing a multispectral compression scheme based on a Karhunen Loeve transform (KLT) followed by the classical DWT+BPE. But such results can be achieved only on perfectly registered bands; a default of registration as low as 0.5 pixel ruins all the benefits of multispectral compression.

In this work, we first study the possibility to implement a multi-bands subpixel onboard registration based on registration grids generated on-the-fly by the satellite attitude control system and simplified resampling and interpolation techniques. Indeed bands registration is usually performed on ground using sophisticated techniques too computationally intensive for onboard use. This fully quantized algorithm is tuned to meet acceptable registration performances within stringent image quality criteria, with the objective of onboard real-time processing. In a second part, we describe a FPGA implementation developed to evaluate the design complexity and, by extrapolation, the data rate achievable on a spacequalified ASIC. Finally, we present the impact of this approach on the processing chain not only onboard but also on ground and the impacts on the design of the instrument.
© (2013) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Mathieu Albinet, Roberto Camarero, Maxime Isnard, Christophe Poulet, and Jokin Perret "Improving multispectral satellite image compression using onboard subpixel registration", Proc. SPIE 8871, Satellite Data Compression, Communications, and Processing IX, 887106 (24 September 2013); https://doi.org/10.1117/12.2024066
Lens.org Logo
CITATIONS
Cited by 1 scholarly publication.
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Image registration

Satellites

Digital signal processing

Image compression

Image quality

Field programmable gate arrays

Satellite imaging

Back to Top