Share Email Print
cover

Proceedings Paper

Real-time 2D floating-point fast Fourier transforms for seeker simulation
Author(s): Richard Chamberlain; Eric Lord; David J. Shand
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The floating point Fast Fourier Transform (FFT) is one of the most useful basic functions available to the image and signal processing engineer allowing many complex and detailed special functions to be implemented more simply in the frequency domain. In the Hardware-in-the-Loop field an image transformed using FFT would allow the designer to think about accurate frequency based simulation of seeker lens effects, motion blur, detector transfer functions and much more. Unfortunately, the transform requires many hundreds of thousands or millions of floating point operations on a single modest sized image making it impractical for realtime Hardware-in-the-Loop systems. .until now. This paper outlines the development, by Nallatech, of an FPGA based IEEE floating point core. It traces the subsequent use of this core to develop a full 256 X 256 FFT and filter process implemented on COTS hardware at frame rates up to 150Hz. This transform can be demonstrated to model optical transfer functions at a far greater accuracy than the current spatial models. Other applications and extensions of this technique will be discussed such as filtering for image tracking algorithms and in the simulation of radar processing in the frequency domain.

Paper Details

Date Published: 12 July 2002
PDF: 9 pages
Proc. SPIE 4717, Technologies for Synthetic Environments: Hardware-in-the-Loop Testing VII, (12 July 2002); doi: 10.1117/12.474723
Show Author Affiliations
Richard Chamberlain, Nallatech Ltd. (United Kingdom)
Eric Lord, Nallatech Ltd. (United Kingdom)
David J. Shand, Nallatech Ltd. (United Kingdom)


Published in SPIE Proceedings Vol. 4717:
Technologies for Synthetic Environments: Hardware-in-the-Loop Testing VII
Robert Lee Murrer, Editor(s)

© SPIE. Terms of Use
Back to Top