On Thu, 26 Apr 2001, Markus Neteler wrote:
thanks for clarification. What do you think: Can we live with that?
Probably an interpolating algorithm would produce smoother results?
But I am not sure what's preferred.
Markus, et al.:
My understanding of interpolation algorithms is that a different one is
most appropriate for different types of data. For example, whether the data
are regularly spaced (as on a grid) or randomly spaced. Also, whether each
measured point represents an actual maxima or minima, or is an estimate.
For example, elevation readings in the Real World are usually randomly
spaced, and each recorded value is probably not a local high- or low-point.
Therefore, an inverse-distance-weighted interpolation algorithm that puts
the fitted line/surface slightly above each maximum point (and slightly
below each minimum point) produces a more accurate estimatation of the true
surface.
On the other hand, if you are interpolating measured chemical values (say,
in soils in a field) along a regular grid, a nearest neighbor interpolation
algorithm that fits the line/surface through each point produces a better
representation of the spatial distribution of measured values.
I've not looked at photo-rectification in more than a couple of years, so
I cannot suggest off-hand just what would be best here. There are, of
course, more interpolation algorithms that ought to be considered, too.
Rich
Dr. Richard B. Shepard, President
Applied Ecosystem Services, Inc. (TM)
2404 SW 22nd Street | Troutdale, OR 97060-1247 | U.S.A.
+ 1 503-667-4517 (voice) | + 1 503-667-8863 (fax) | rshepard@appl-ecosys.com
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'