Distilling Noise Sources for Digital Capture Devices

A procedure that enables the evaluation and further distillation of fixed pattern noise for digital capture devices is proposed. This method also allows for the direct calculation of the random, or temporal, noise contribution, devoid of fixed pattern components. Our method is consistent with a draft digital camera noise measurement standard under consideration through ISO 15739, but is intended to be a more general approach to the problem. Using common measurement protocols for statistical estimation, a demonstration of the technique for a desktop reflection scanner is provided, accompanied by data analysis. This analysis showed that over 90% of the variance was due to fixed pattern noise. Surprisingly, over 75% was due to the target's microstructure and not the scanner itself. Applications for this method include device performance verification, engineering system analysis, and target noise specification.