Recall the Seabeam data of Figure
.
In chapter
we filled empty bins minimizing the output of Laplace's operator,
getting Figure
.
The problem with the Laplacian operator as an interpolator
is that it smears information uniformly in all directions.
From Figure
we see
that we need an anisotropic interpolation
oriented along the regional trends.
What we need is a PEF in place of the Laplacian.
To get it,
we apply module pef
.
After binning the data and finding this PEF,
we do a second stage of linear-least-squares optimization
with mis2
, as we did for Figure
14,
and
we obtain the pleasing result in Figure 20.
![]() |