Surface profile tolerance measuring method using kriging method

The measuring time required to inspect a machined part using a coordinate measuring machine (CMM), especially for a given form or profile tolerance, should be minimized and this can be achieved by reducing the number of sample points obtained from a machined surface. In this paper, variogram and kriging models, both well-known, are evaluated in terms of the degrees of prediction accuracy given in an NURBS surface and machined surface. The Gaussian variogram model and ordinary kriging model are utilized to predict the machined surface profiles. For faster inspection, the number of sample data is reduced by using a modified median filtering based on both models. Finally, both maximum and minimum points on a machined surface are searched in order to inspect a given tolerance range by using the geometric similarity and a sphere. The research results presented here will be valuable in enabling the use of a 3-axis CMM rather than the more expensive 5-axis CMM.