DocumentCode
1923475
Title
Deconvolution of VNIR spectra using modified Gaussian modeling (MGM) with automatic parameter initialization (API) applied to CRISM
Author
Makarewicz, Heather D. ; Parente, Mario ; Bishop, Janice L.
Author_Institution
EECS & Math Depts., Univ. of Kansas, Lawrence, KS, USA
fYear
2009
fDate
26-28 Aug. 2009
Firstpage
1
Lastpage
5
Abstract
Reflectance spectroscopy is a powerful tool for determining mineralogy in both laboratory and field experiments. Several studies indicate that reflectance spectra can be modeled as a sum of modified Gaussian functions and a continuum, which is called the modified Gaussian model (MGM). In this study, a method for automatic parameter initialization (API) for the MGM is proposed that is based solely on the spectrum being modeled. The API determines the number of Gaussians to model and their initial parameter estimates. The MGM with API has been tested with artificial, laboratory, and CRISM spectra. Initial results indicate that the method is successful on a variety of mineral spectra.
Keywords
Gaussian distribution; curve fitting; feature extraction; geophysical signal processing; minerals; remote sensing; splines (mathematics); Gaussian distributions; automatic parameter initialization; curve fitting; mineralogy; modified Gaussian functions; modified Gaussian model; reflectance spectroscopy; remote sensing; Deconvolution; Laboratories; Light sources; Mars; Minerals; NASA; Optical imaging; Reconnaissance; Reflectivity; Spectroscopy; Curve fitting; Gaussian distributions; Remote Sensing; Spectroscopy;
fLanguage
English
Publisher
ieee
Conference_Titel
Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, 2009. WHISPERS '09. First Workshop on
Conference_Location
Grenoble
Print_ISBN
978-1-4244-4686-5
Electronic_ISBN
978-1-4244-4687-2
Type
conf
DOI
10.1109/WHISPERS.2009.5289046
Filename
5289046
Link To Document