Logo of the Physikalisch-Technische Bundesanstalt
symbolic picture: "magazines"

Enhanced spectral color measurement procedure

Within the scope of a cooperation project with industry, PTB has developed a novel approach to the quality control of color spectra

PTB-News 1.2017
Especially interesting for

spectral data analysis

multispectral image acquisition systems

industrial color inspection

Today, spectral color measurements are indispensable in numerous areas of science and industry. In the printing industry, for example, clear color definitions and exact color repetitions are necessary. The procedures used to date, however, only provide random and locally limited results – and some require a large amount of time and effort. Within the scope of a joint project with the Chromasens GmbH company, PTB has developed a procedure that allows color spectra to be realized with great efficiency and accuracy based on prior spectral information. A simulation study has demonstrated that this approach provides considerably better results than conventional procedures.

The multispectral line scan camera decomposes the color spectrum. (Photo: Chromasens GmbH)

color spectrum.

Each color has its individual “finger-print”, the color spectrum which consists of different wavelengths and intensities of light. Measuring the spectrum discloses which fractions of the light are reflected and reach the observer – it is therefore called a “reflectance spectrum”. Standard procedures that are currently used for spectral color measurement determine the reflectance spectrum only punctually.

Thus, large surfaces cannot be detected as a whole. Moreover, such a measurement is tedious: when checking the quality of printing, for example, a manual comparison between the target colors and the actual colors is carried out by means of special color measuring in-struments on test sheets.

In the future, this step might become redundant. By means of a multispectral line scan camera and of an algorithm which was developed at PTB, it will be possible to record – fully automatically and in real time – the full reflectance spectrum of a surface to be measured. The camera used is equipped with 12 color channels which, similar to the cones in the human eye, react to light with a different sensitivity. First, the spectral sensitivity curves of the individual channels are determined. This is the basis of the subsequent reconstruction of the color spectrum from the measured data of the camera. The recon-struction procedure is based on Bayesian statistics. Hereby, the measurement results of the camera are linked up with previous knowledge on the spectrum. This approach allows the reli-able characterization of the quality of the measurement results.

A simulation study has demonstrated that the new measurement procedure yields results that are much more reliable than those obtained by means of the conventional, empirical procedure. PTB's project with the Chromasens GmbH company has been extended for another two years. The objective consists in paving the way for a wide use of this new method. Due to the acquisition of spectral properties of sample surfaces, it is thus possible to perform efficient material sorting and conformity checks in industrial production.


Marcel Dierl
Department 8.4 Mathematical Modelling and Data Analysis
Phone: +49 (0)30 3481-7978

Scientific publication

M. Dierl, T. Eckhard, B. Frei, M. Klammer, S. Eichstädt, C. Elster: Improved estimation of reflectance spectra by utilizing prior knowledge. J. Opt. Soc. Am. A. 33, 1370−1376 (2016)