Computational geometry in the preprocessing of point clouds for surface modeling

In Computer Aided Geometric Design ( CAGD ) the automated fitting of surfaces to massive series of data points presents several difficulties: (i) even the formal definition of the problem is ambiguous because the mathematical characteristics (continuity, for example) of the surface fit are dependent on non-geometric considerations, (ii) the data has an stochastic sampling component that cannot be taken as literal, and, (iii) digitization characteristics, such as sampling interval and directions are not constant, etc -- In response, this investigation presents a set of computational tools to reduce, organize and re-sample the data set to fit the surface -- The routines have been implemented to be portable across modeling or CAD servers -- A case study is presented from the footwear industry, successfully allowing the preparation of a foreign, neutral laser digitization of a last for fitting a B-spline surface to it -- Such a result was in the past attainable only by using proprietary software, produced by the same maker of the digitizing hardware