Adaptive thinning for bivariate scattered data (Q1612421)

The authors derive algorithms for constructing significant point sets from large sets of scattered data points. Algorithms of this kind are important in many fields of applications ranging from fast visualization to interpolation on scattered point sets in oil field exploration. The thinning of the large point set is done adaptively in the sense that the error of piecewise linear interpolation on a Delaunay triangulation of the smaller subsets is small. Numerical tests are provided to demonstrate the ability of the derived algorithms.