Inline color measurement systems measure color on the press while the web is moving, allowing press operators to measure color continuously throughout a run. The amount of data generated by inline color measurement systems exceeds offline color measurement systems and must be analyzed using different methods than used for analyzing offline color measurement data. The statistical process monitoring (SPM) methods for analyzing inline color measurement data must take into account the units of measure, amount of data, and organization of the data.
Color difference data is best suited for monitoring printed color in relation to tolerances set by the printer or print buyer, but is less suited for statistical analysis because it relates little to the physical aspects of the printed color. Plots of CIE DL*Da*Db* and CIE DL*DC*Dh are best suited for informing the printer about how printed color is different from a Standard L*a*b* target, but cannot be used to monitor color relative to DE tolerances. Spectral reflectance is a direct physical measurement of the printed process, and using Principle Components Analysis, can be decomposed into two or three orthonormal eigenvectors that can be monitored independently using SPM techniques.
Regardless of the color measurement units, the SPM method must be carefully considered. Commonly used control charts, such as I/MR and x/R, assume independence and normality the data, assumptions generally not valid for printing processes where features such as process drifting and process shifts are common, thus requiring the use of more advanced SPM methods.