Basic concept of feature based metrology

Abstract

In coordinate metrology, an associated feature (Gaussian associated feature) is normally calculated from an extracted feature which is determined by a measured data set of CMM (Coordinate Measuring Machine) using a least squares method. This data processing flow which is called as "feature based metrology" disagrees with the data processing methods in profile metrology and length measurement. The most significant difference between feature based metrology and profile metrology is quantity of a number of measured points on each measured feature. In feature based metrology, the number of measured points is very small and a data sampling strategy is a discrete sampling. On the other hand, a continuous sampling strategy to measure many continuous points is adapted to profile metrology. In this paper, the basic concept of feature based metrology is discussed in defining a feature model, calculating parameters of feature and estimating uncertainty of measurement. Three-dimensional modeling for three-dimensional feature, least squares method and statistical estimation strategy for estimating the uncertainty of the feature agree with feature based metrology. The series of simulations for the feature based metrology in statistical way directly implies that the basic concept and the basic data processing method in this paper are useful to feature based metrology.