1 January 2010 Analysis and assessment of the effects of fixed pattern and quantization noise on the accuracy of color rendition in wide-dynamic-range complementary metal-oxide semiconductor imagers
Stephen Otim, Stephen Collins
Author Affiliations +
Abstract
For wide-dynamic-range cameras that achieve a logarithmic response using a load transistor operating in weak inversion, fixed pattern noise is a particular problem. The effects of fixed pattern noise can be reduced using one of a variety of methods that compensate for different forms of fixed pattern noise. In choosing between these methods, it becomes important to know the level of fixed pattern noise that is tolerable for the application in question. Similarly, when deciding on the number of bits that should be used when digitizing the image, it is important to know the level of quantization noise that is tolerable to maintain color image quality. In this work, a method of determining the tolerable levels of fixed pattern and quantization noise is proposed. This leads to the conclusion that for one type of wide-dynamic-range pixel, a relatively simple fixed pattern noise correction procedure gives acceptable results for applications of medium complexity. A more generally applicable result is that if quantization levels equivalent to a contrast change of less than 2% in logarithmic imagers are obtainable, then excellent color image rendition can be achieved.
©(2010) Society of Photo-Optical Instrumentation Engineers (SPIE)
Stephen Otim and Stephen Collins "Analysis and assessment of the effects of fixed pattern and quantization noise on the accuracy of color rendition in wide-dynamic-range complementary metal-oxide semiconductor imagers," Journal of Electronic Imaging 19(1), 011011 (1 January 2010). https://doi.org/10.1117/1.3272597
Published: 1 January 2010
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Quantization

Image quality

Imaging systems

Cameras

Transistors

Error analysis

RGB color model

Back to Top