Effect of void pixels on the quantification of surface topography parameters

In the last decades the measurement of surface topographies was further developed using the capabilities of 3D optical profilers based on different measuring principles. These instruments have significant advantages over contact stylus measuring instruments but also some limitations. One limitation is depending from the interaction of the surface with the numerical aperture of the objective. Because of the surface slope, some parts of the surface are not correctly detected by the instrument, resulting on void pixels i.e. missing information on the measured topography. Data generated using fitting algorithms can be used to reconstruct the surface and replace the missing data, and the way this operation is performed affects the evaluation of surface parameters. As an alternative, the calculation of parameters may be performed on non-reconstructed datasets. In this work, the effect of void pixels on the determination of 3D parameters on different manufactured surfaces has been investigated using computer simulation starting from actual measurement data. Both randomly distributed and slope-dependent void pixels are considered. A sensitivity analysis with different distributions of void pixels as well as the effect of increasing slope are also presented.