Chapter 3 - Image filtering and morphology

Abstract

This chapter starts by discussing the use of linear filters and then goes on to deal with nonlinear filters, including median, mode, and rank order filters. It finds that median filters have a firm role in removing impulse noise, while mode filters are better adapted to image enhancement. This chapter explores the incidental shifts made by median and other filters. While these shifts normally represent undesirable distortions, in the cases of min and max filters they are actually useful. This chapter covers dilation, erosion, opening, and closing, and presents a number of theorems describing the properties of such operators. The value of residue operators for locating various features and defects is discussed. The concepts are generalized to grayscale processing, and it is shown how morphological operations can be used to group particles. Overall, the chapter emphasizes how morphology provides a firmer theoretical foundation for shape analysis, integrating a range of disparate topics.