Comparison of two free-form surfaces based on discrete data points is
of paramount importance for reverse engineering. It can be used to ass
ess the accuracy of the reconstructed surfaces and to quantify the dif
ference between two such surfaces. The entire process involves three m
ain steps: data acquisition, 3D feature localization and quantitative
comparison. This paper presents models and algorithms for 3D feature l
ocalization and quantitative comparison. Complex free-form surfaces ar
e represented by bicubic parametric spline surfaces using discrete poi
nts. A simple yet effective pseudoinverse algorithm was developed and
implemented for localization. It consists of two iterative operations,
namely, constructing a pseudo transformation matrix and point matchin
g. A computing algorithm was developed to compare two such surfaces us
ing optimization techniques. Since this approach does not involve solv
ing non-linear equations for the parameters of positions and orientati
ons, it is fast and robust. The algorithm was implemented and tested w
ith several examples. It is effective and can be used in industry for
sculptured surface comparison. Copyright (C) 1996 Published by Elsevie
r Science Ltd