Share Email Print

Proceedings Paper

Spectral estimation and processing of imagery data
Author(s): David H. Tofsted
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Given the task of modifying color images for atmospheric effects one step involves recasting the color (red, green, blue) pixel information as spectrally resolved information that can be modified for transmission, path radiance, and aerosol-forward-scatter/turbulence blurring effects. The paper describes the means of using sensor spectral response functions to perform the spectral estimation task. Spectral data is recorded as a function of pixel position and waveband using a series of object-oriented C++ classes. Each waveband image is processed separately for atmospheric effects, and the output spectral images are used to create an output color image. The estimation technique begins by defining a color space that is the analog of human chromaticity space. Sensor spectral responsivity data is required as input and used in providing a series of monoband response points and three templates related to the red, green, and blue response functions. The estimation technique involves direct calculations for color coordinates within the red/green/blue triangle. For valid color coordinates outside this triangle, an iterative technique was developed.

Paper Details

Date Published: 15 June 1994
PDF: 10 pages
Proc. SPIE 2223, Characterization and Propagation of Sources and Backgrounds, (15 June 1994); doi: 10.1117/12.177941
Show Author Affiliations
David H. Tofsted, U.S. Army Research Lab. (United States)

Published in SPIE Proceedings Vol. 2223:
Characterization and Propagation of Sources and Backgrounds
Wendell R. Watkins; Dieter Clement, Editor(s)

© SPIE. Terms of Use
Back to Top