Multi-spectral imaging and spectral reflectance reconstruction can be used in cultural-heritage institutes to digitalize their collections for documentation purposes. It can be used to simulate artwork under any lighting condition, and to analyze colorants that were used. The basic idea of a multi-spectral imaging system is to sub-sample spectral reflectance factor, producing results similar to a spectrophotometer. The sampled data are used to reconstruct reflectance for the visible spectrum.

In this thesis, a wide band multispectral camera was designed and constructed to achieve high spectral and color accuracy as well as high image quality.

Noise propagation theory was introduced and tested. A seven channel band- pass filter set was modeled using Gaussian functions and optimized to yield high spectral and colorimetric reproduction accuracy as well as low colori- metric noise. Single and "sandwich" filters were selected from o!-the-shelf absorption filters using the Gaussian bandpass filter model. Experiments were conducted to test the spectral, color and noise performance of the novel sandwich filters and compared with interference filters. The novel sandwich fil- ters led to increased colorimetric accuracy along with a reduction colorimetric noise.

This imaging system will be used as part of a recommended workflow for museum archiving, and will be an important addition to the spectral imaging capabilities at MCSL.

Library of Congress Subject Headings

Multispectral imaging--Equipment and supplies; Colorimetry; Cameras--Design and construction

Publication Date


Document Type


Student Type


Degree Name

Color Science (MS)


Roy S. Berns

Advisor/Committee Member

Mark D. Fairchild


Physical copy available from RIT's Wallace Library at TR267.733.M85 W36 2016


RIT – Main Campus

Plan Codes