Abstract

Fluorescent effects have been observed for thousands of years. Stokes, in 1852, began the science of fluorescence culminating in his law of fluorescence, which explained that fluorescence emission occurs at longer wavelengths than the excitation wavelength. This phenomenon is observed extensively in the art world. Daylight fluorescent colors known as Day-Glo have become an artistic medium since the 1960s. Modern artists exploit these saturated and brilliant colors to glitter their painting. Multispectral imaging as a noninvasive technique has been used for archiving by museums and cultural-heritage institutions for about a decade. The complex fluorescence phenomenon has been often ignored in the multispectral projects. The ignored fluorescence results in errors in digital imaging of artwork containing fluorescent colors. The illuminant-dependency of the fluorescence radiance makes the fluorescence colorimetry and consequently spectral imaging more complex. In this dissertation an abridged imaging bi-spectrometer for artwork containing both fluorescent and non-fluorescent colors was developed. The method developed included two stages of reconstruction of the spectral reflected radiance factor and prediction of the fluorescent radiance factor. The estimation of the reflected radiance factor as a light source independent component was achieved by imaging with a series of short-wavelength cutoff filters placed in the illumination path. The fluorescent radiance factor, a light source dependent component, was estimated based on a proposed model, the abridged two-monochromator method. The abridged two-monochromator method was developed for reconstructing the bi-spectral matrix of a fluorescent color based on a calibrated UV-fluorescence imaging. In this way, one could predict the fluorescence radiance factor under any desired illuminant and consequently a better color evaluation and rendering could be obtained. Furthermore, this method easily fitted in a general system for spectral imaging of paintings containing both fluorescent and non-fluorescent colors. The abridged two-monochromator method could predict fluorescent radiance factor of a fluorescent color via prediction of the true emission and the number of absorbed quanta by a fluorescing specimen for a given viewing light source. The superiority of the abridged fluorescence spectral imaging to the traditional spectral and colorimetric imaging for a few light sources was confirmed using fluorescent and non-fluorescent targets. Additionally, an exploratory visual experiment using a paired-comparison method was performed to evaluate the performance of the abridged fluorescence spectral imaging in comparison to the traditional spectral and colorimetric imaging for rendering images of a reference painting. The abridged fluorescence spectral imaging had better performance than traditional spectral and colorimetric imaging in rendering images for daylight.

Library of Congress Subject Headings

Fluorescence spectroscopy; Spectrum analysis; Colorimetry

Publication Date

2-23-2009

Document Type

Dissertation

Student Type

Graduate

Department, Program, or Center

Chester F. Carlson Center for Imaging Science (COS)

Advisor

Berns, Roy

Advisor/Committee Member

Fairchild, Mark

Advisor/Committee Member

Langner, Andreas

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: QD96.F56 M64 2009

Campus

RIT – Main Campus

Share

COinS