| Re: matching irregular data sets [message #27908 is a reply to message #27786] |
Thu, 08 November 2001 17:25   |
air_jlin
Messages: 22
Registered: July 2001
|
Junior Member |
|
|
hi Steve,
i'm not an idl guru, but my experience with interpolating large fields
is that it's just plain slow (i've tried a couple methods).
if you're interested in how i did my near neighbor, you can see it at:
http://www.johnny-lin.com/lib.html#atmos
it's called NN_INTERP, and isn't the prettiest, but it seems to work.
the main slow down is that it calculates the distances between *all*
possible location pairs, since i wanted the procedure to be able to
accomodate irregular input and output grids. if you have a timeseries
of fixed grids, you can speed things up tremendously by precalculating
the distances (the procedure has that option in it).
hope this helps!
best,
-Johnny
-------------------------------------------
Johnny Lin
CIRES, University of Colorado
Work Phone: (303) 735-1636
Web: http://cires.colorado.edu/~johnny/
-------------------------------------------
"Steve W. Nesbitt" <snesbitt@met.utah.edu> wrote in message
news:<3BE9C7B3.DE76D102@met.utah.edu>...
> Howdy,
>
> This may or may not be a dumb question. I am working on matching two
> irregularly-spaced remote sensing data sets, specifically doing a
> nearest-neighbor or bilinear interpolation of one data set to another.
> I have written a routine to do this, but it is painfully slow since the
> arrays I'm matching are lat/lon grids [400,3000]. I have searched the
> manual ad nauseum! for an IDL canned routine to do this, but they seem
> to require that the output grids be regularly spaced. I would like the
> output to be gridded to the second irregular grid, and it would be nice
> if it would return the indices of the original grid in the output. Let
> me know if one of you IDL gurus can help me out on this one.
>
> Many thanks,
> -Steve
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
