Fastgrid is a R package for gridding point data to an uniform grid. It uses Gaussian process regression (aka Kriging) with exponential covariance function. It only calculates point predictions (no uncertainties) and tries to be as efficient as possible. The loop over grid points is done in Fortran using OpenMP parallelization. Interpolation to 1600 x 400 grid takes about 15 seconds.
In the example there are some references to plotting the results by some code that I have not yet provided, sorry.
Marko Laine, [email protected]