The still imaging portion of FADGI [1] continues to be a living document that has evolved from its theoretical digital imaging principles of a decade ago into adaptations for the realities of day-to-day cultural heritage workflows. While the initial document was a bit disjointed, the 2016 version is a solid major improvement and has proven very useful in gauging digital imaging goodness. [2] With coaching, encouragement and focused attention to detail many users, even the unschooled, have achieved 3-star compliance, sometimes with high-speed sheet-fed document scanners. 4-star levels are not far behind. This is a testimony to an improved digital image literacy for the cultural heritage sector that the authors articulated at the beginning of the last decade. This objective and science based literacy has certainly evolved and continues to do so. It is fair to say that no other imaging sector has such comprehensive objective imaging guidelines as those of FADGI, especially in the context of high volume imaging workflows. While initial efforts focused on single instance device benchmarking, future work will concentrate on performance consistency over the long term. Image digitization for cultural heritage will take on a decidedly industrial tone. With practice, we continue to learn and refine the practical application of FADGI guidelines in the preservation of meaningful information. Like rocks in a farm field, every year new issues and errors with current practices surface that were previously hidden from view. Some are incidental, others need short term resolution. The goal of this paper is to highlight these and make proposals for easier, less costly, and less frustrating ways to improve imaging goodness through the FADGI guidelines.
This paper presents two on-chip calibration methods for improving the linearity of a CMOS image sensor (CIS). A prototype 128 × 128 pixel sensor with a size of 10 μm×12 μm is fabricated using a 0.18 μm 1P4M CIS process. Both calibration methods show obvious improvement on the linearity of the CIS. Compared with the voltage mode (VM) calibration, the pixel mode (PM) calibration method achieves better linearity results by improving the nonlinearity of the CIS 26×. This results in a minimum nonlinearity of 0.026%, which is a 2× better than the state-of-the-art.