Surface characterization by morphological filtering

Morphological filters are non-linear methods that developed from the transfer of set-theoretic operations to image analysis. In the context of surfaces, they yield an adequate description of the sampling effect of both mechanical styli and other probing systems such as Scanning Probe Microscopy. The paper illustrates the fundamentals of morphological filters, discusses their advantages over conventional linear filters, and focuses on their application to surface characterization. The problem of recovering the true 3D profile is addressed, and the requisites to prevent a loss of information are discussed. Moreover, new shape-oriented methods to extract form, waviness, and roughness from profile data are presented.