Chapter 7 - Image Noise Models

Abstract

Noise occurs in images for many reasons. Probably the most frequently occurring noise is additive Gaussian noise. It is widely used to model thermal noise and, under some often reasonable conditions, is the limiting behavior of other noises, e.g., photon counting noise and film grain noise. Gaussian noise is a part of almost any signal. For example, the familiar white noise on a weak television station is well modeled as Gaussian. Since image sensors must count photons—especially in low-light situations—and the number of photons counted is a random quantity, images often have photon counting noise. The grain noise in photographic films is sometimes modeled as Gaussian and sometimes as Poisson. Many images are corrupted by salt and pepper noise, as if someone had sprinkled black and white dots on the image. Other noises include quantization noise and speckle in coherent light situations.