nn
(Natural Neighbours interpolation)
nn is a C code for Natural Neighbours interpolation of 2D scattered data. It provides a C library and a command line utility nnbathy.
Algorithmically, it was initially loosely based on the Dave Watson's description of nngridr; code-wise it is an independent development. You may see a comparison of performance of a (rather old) version of nn with nngridr in
Quanfu Fan, Alon Efrat, Vladlen Koltun, Shankar Krishnan, and Suresh Venkatasubramanian. Hardware-assisted Natural Neighbor Interpolation. In Proc. 7th Workshop on Algorithm Engineering and Experiments (ALENEX), 2005. pdf
nn is coded for robustness (to handle degenerate data) and scalability (to handle millions of data points), subject to using double precision calculations. For the underlying Delaunay triangulation it calls exact arithmetic code from triangle.
Checkout nn by running git clone https://github.com/sakov/nn-c
or svn checkout https://github.com/sakov/nn-c
.