Fitting parametrized polynomials with scattered surface data

Currently used joint-surface models require the measurements to be structur ed according to a grid. With the currently available tracking devices a lar ge quantity of unstructured surface points can be measured in a relatively short time. In this paper a method is presented to fit polynomial functions to three-dimensional unstructured data points. Tea test the method spheric al, cylindrical, parabolic, hyperbolic, exponential, logarithmic, and sella r surfaces with different undulations were used. The resulting polynomials were compared with the original shapes. The results show that even complex joint surfaces can be modelled with polynomial functions. In addition, the influence of noise and the number of data points was also analyzed. From a surface (diam: 20 mm) which is measured with a precision of 0.2 mm a model can be constructed with a precision of 0.02 mm. (C) 1999 Elsevier Science L td. All rights reserved.