Abstract
A Quad Bayer pixel structure is the current driving technology in developing high resolution image sensors with smaller pixel pitch. However, by reducing the pixel size the signal-to-noise ratio (SNR) and the dynamic range (DR) of an image sensor will be compromised. One of the means to improve the SNR and DR for a quad-cell image sensor or other type of complementary metal-oxide semiconductor (CMOS) image sensor with tiny pixels is to use a dual conversion gain (DCG) approach in a pixel structure where the conversion gain (CG) selection is adaptive with scene illumination and enhances the SNR and DR. However, current DCG image sensors use the same CG for all of the pixels within one frame (or one image) to obviate nonuniformity pattern artifacts that would appear within an image (because of the pixel dark offset values and also the light response that may be changed with a CG). To address the challenge that was described above, techniques are presented herein that support two-pixel readout mechanisms and image calibration means for a full resolution (i.e., remosaicing) mode of quad-cell image sensors that enable each individual pixel’s gain to be selected independently from the remaining pixels. Such unique readout mechanisms sense the exposure level of each pixel, set the conversion gain of each pixel individually in real-time, and extend analog-to-digital convertor (ADC) bit depth by adding gain bits to achieve the highest DR and SNR. Further, each pixel may be characterized by a unique dark and flat field calibration for each CG setting. Still further, dark and flat field calibrations may be applied to correct any nonuniformity pattern image artifacts that may result from the use of multiple CGs within a single image. Aspects of the presented techniques support image calibration and correction activities. Importantly, the presented techniques may be extended for all types of image sensors.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Gholmansaraei, Farhad Abbassi, "PIXEL-BASED CONVERSION GAIN SELECTION FOR QUAD-CELL CMOS IMAGE SENSORS WITH IMAGE CALIBRATION AND EXTENDED DYNAMIC RANGE", Technical Disclosure Commons, (October 05, 2022)
https://www.tdcommons.org/dpubs_series/5420