Home »
Public Forums »
archive »
Management of griddata around the borders of the data
| Management of griddata around the borders of the data [message #50035] |
Tue, 05 September 2006 11:37 |
Matt Harner
Messages: 5
Registered: September 2006
|
Junior Member |
|
|
> From what we can read, there is no way to manage border/missing data
using a keyword in GRIDDATA. From what we can read, the only way to
manage GRIDDATA around borders is to explicitly eliminate the border
data from the input. From what we can read, the TRIANGULATE function
has a B keyword that identifies the border pixels in some encoded
fashion. 1) Do we assume that GRID_INPUT can take the B output from
TRIANGULATE in the EXCLUDE keyword? 2) Will this fix the problem at
the borders?
BTW the problem at the borders is that the GRIDDATA does a very bad job
of interpolation on some of the edge values. It returns a value much
larger or much smaller than the minimum or maximum of our original
data.
We determined that this is a problem, by simply removing those very
large or very small values. The remaining data looked more
appropriate. We do not know about intermediate values which may also
be incorrect. Hence we need a more systematic way of determining which
data points to exclude.
3) What is a systematic way to determine which data points will be
interpolated incorrectly and hence which data points to exclude from
the GRIDDATA input?
So far, we have been using the linear interpolation option. We would
prefer to use the cubic or bi-cubic interpolation.
4) Will the answer to the GRIDDATA problem also help with the other
interpolation methods?
|
|
|
|
Current Time: Sat Oct 11 06:16:36 PDT 2025
Total time taken to generate the page: 1.84031 seconds
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
