Lossless Compression for Space Imagery in a Dynamically Reconfigurable Architecture Xiaolin Chen 1, , C. Nishan Canagarajah 1 , Raffaele Vitulli 2 , and Jose L. Nunez-Yanez 1 1 Department of Electrical and Electronic Engineering, University of Bristol, UK {Xiaolin.Chen,Nishan.Canagarajah,J.L.Nunez-Yanez}@bristol.ac.uk 2 European Space Agency (ESA), On-Board Payload Data Processing Section, The Netherlands Raffaele.Vitulli@esa.int Abstract. This paper presents a novel dynamically reconfigurable hard- ware architecture for lossless compression and its optimization for space imagery. The proposed system makes use of reconfiguration to support optimal modeling strategies adaptively for data with different dimen- sions. The advantage of the proposed system is the efficient combination of different compression functions. For image data, we propose a new multi-mode image model which can detect the local features of the im- age and use different modes to encode regions with different features. Experimental results show that our system improves compression ratios of space image while maintaining low complexity and high throughput. 1 Introduction Advances in remote sensing facilities generate massive amounts of data. For in- stance, SAR (Synthetic Aperture Radar) and the LANDSAT can produce hun- dreds of gigabytes of data per day, not to mention the 224 bands hyperspectral image from AVIRIS (Airborne Visible/Infrared Imaging Spectrometer). These data are transmitted to the Earth for further processing. However, the data vol- ume is often several times larger than the transmission capacity of the downlink circuit, which limits the scientific data return from the spaceborne instruments. This is known as the “Bandwidth vs. Data Volume” challenge for modern space- craft [1]. An effective solution to this problem is compression. As space data is costly and subject to processing, all the information in the data should be preserved. Therefore, lossless compression is necessary in space applications. There are various kinds of data that need to be sent from spacecraft to the Earth, such as 1-D general data, 2-D image data, 3-D multispectral image data or video, etc. These data are likely to be transmitted along the same physical channel. Therefore, a system that can compress different types of data with real- time adaptation is of interest to space applications. The Consultative Committee for Space Data Systems (CCSDS) recommended an ASIC device PRDC (Payload  Thanks to the support from EPSRC under grant EP/D011639/1. R. Woods et al. (Eds.): ARC 2008, LNCS 4943, pp. 336–341, 2008. c  Springer-Verlag Berlin Heidelberg 2008