Specifications

Submitted to Studies in Conservation, March 2006

Practical Spectral Imaging Using a Color-Filter Array Digital

Camera

Roy S. Berns

Lawrence A. Taplin

Mahdi Nezamabadi

Yonghui Zhao

Mahnaz Mohammadi

Munsell Color Science Laboratory

Chester F. Carlson Center for Imaging Science

Rochester Institute of Technology

54 Lomb Memorial Drive

Rochester, NY 14623-5604, USA

Summary

Spectral-based imaging facilitates image archives with high colorimetric accuracy and the

opportunity for quantitative analysis, in similar fashion to spectral-based analytical techniques

common to conversation science. Most commonly, such systems are imaging spectrometers,

sampling the visible spectrum at between 10 and 50 nm bandwidth and interval. They are

complex, expensive, and require considerable imaging science expertise. This publication

describes an alternate approach that results in a practical system, appropriate for museum,

libraries, and archives. A professional-grade, color-filter-array digital camera was modified by

removing its infrared cover glass and replacing it with clear glass. Two filters, blue-green and

yellow, were designed for placement in the optical path sequentially. Design criteria included

spectral and colorimetric accuracy, image noise, capture time, ultraviolet and infrared radiation

rejection, and fabrication simplicity and cost. The pair of images were registered and corrected

for dark noise and spatial inhomogeneities. Using a calibration target of colored samples with

known optical properties as a function of wavelength, a transformation was derived that

converted camera signals to spectral reflectance factor. Deriving the transformation matrix was

based on the Wyszecki hypothesis in which a spectrum can be decomposed into a fundamental

stimulus (defining its color) and a metameric black (defining its colorants). The system was