Pixels are the elementary units in digital image processing. A number of image processing tasks can be performed by simply handling these pixels as individual objects or measuring points. As long as the statistical properties of a pixel do not depend on its neighbor pixels, we can apply the classical concepts of first-order statistics which are used to handle measurements of single scalar quantities, e. g., temperature, pressure, and size. The random nature of the signal measured at individual pixels can be related to one or more of the following processes:

1. The imaging sensor introduces various types of noise into the measured irradiance. For most — but not all — noise sources, the individual pixels are not correlated to each other and can thus be treated individually.