ABSTRACT
Fixed pattern noise arises in an array of optical detectors due to variability in the response of each detector to uniform stimulus. It appears as a salt-and-pepper pattern on images acquired by the array and hence degrades their image quality. This entry explores the sources of fixed pattern noise in various optical sensors. To improve the image quality, correction of this noise is carried out by determining a parametric model for the individual detectors in the array. Correction of one or a few dominant parameters is often sufficient to obtain good-quality images. This entry describes various parametric correction techniques, including double sampling techniques used in a complementary metal oxide semiconductor active pixel sensors. Furthermore, continuous fixed pattern noise correction techniques based on statistical properties of the images themselves or motion between images are also described.
