Multidimensional set-theoretic feature recognition

Abstract

A new method of feature recognition in set-theoretic geometric modelling is proposed which allows its users complete control over the geometrical definition of form features. The method looks for instances where one set-theoretic model under multidimensional transformations coincides with all of, or part of, another. The attributes of the possible match locations are then reported back to the user. The method also permits the implementation of a tolerance level which allows matches whose goodness of fit is below this level to be ignored. All possible instances of features whose exact geometries are unknown can be defined parametrically. The method reports back the parameter values of match positions along with the other match attributes. The method also has applications to the nesting and docking problems.