Analysis of modified discretization methods for the measurement volume

Abstract

The analysis performed in the paper shows that the effectiveness of discretization methods depends on the accuracy of the evaluation of the parameters of local surface errors and on the characteristics of the regression polynomial describing them. It is evident from the expressions derived that the wavelength of the distribution errors depends on the number of members of the regression polynomial. By increasing the number of members of the regression polynomial, the wavelength of errors of the surface form will be not evaluated. On the other hand, reducing the number of polynomial members, the accuracy of the description of local surface errors will be lost. This is why a priori information is needed about the surface to be measured before choosing the order of the polynomial equation.