Re: Best Fit Plane - Parallel to xyplane!


matt271829-news@xxxxxxxxxxx wrote:

....

The approximation will be
reasonable provided the surface doesn't deviate too extremely from the
square tops of the columns - or deviates in such a way that the
overestimations more-or-less cancel out the underestimations.

I should mention also that you don't necessarily *have* to have an
evenly-spaced grid. If some areas of the surface are more feature-rich
than others then you could pack the sample points more densely in those
regions (say half-spacing, or quarter-spacing), and weight the z-values
according to the cross-sectional area of the columns...

.