2 November 1999 High-performance signal and image processing for remote sensing using reconfigurable computers
Author Affiliations +
It is not uncommon for remote sensing systems to produce in excess of 100 Mbytes/sec. Los Alamos National Laboratory designed a reconfigurable computer to tackle the signal and image processing challenges of high bandwidth sensors. Reconfigurable computing, based on field programmable gate arrays, offers ten to one hundred times the performance of traditional microprocessors for certain algorithms. This paper discusses the architecture of the computer and the source of performance gains, as well as an example application. The calculation of multiple matched filters applied to multispectral imagery, showing a performance advantage of forty- five over Pentium II (450 MHz), is presented as an exemplar of algorithms appropriate for this technology.
© (1999) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Michael P. Caffrey, Michael P. Caffrey, John J. Szymanski, John J. Szymanski, A. Begtrup, A. Begtrup, J. Layne, J. Layne, T. Nelson, T. Nelson, S. Robinson, S. Robinson, Alfonso Salazar, Alfonso Salazar, James P. Theiler, James P. Theiler, } "High-performance signal and image processing for remote sensing using reconfigurable computers", Proc. SPIE 3807, Advanced Signal Processing Algorithms, Architectures, and Implementations IX, (2 November 1999); doi: 10.1117/12.367629; https://doi.org/10.1117/12.367629

Back to Top