| Re: interpolation question [message #48454 is a reply to message #25895] |
Mon, 24 April 2006 12:47   |
news.verizon.net
Messages: 47
Registered: August 2003
|
Member |
|
|
> I am so wonder that why IDL has no simple function like MATLAB's
> 'interp2'.
I agree with you that IDL appears deficient to MATLAB in providing an
easy and consistent set of interpolation routines. First of all,
while IDL has interpol.pro for 1d interpolation, there is no equivalent
function for 2-d interpolation. (bilinear.pro and interpolate
require you to supply indicies). And while interpol.pro provides
several interpolation methods, it doesn't include the simplest (nearest
neighbor), though this would be easy to add (as in
http://idlastro.gsfc.nasa.gov/ftp/pro/math/linterp.pro). It makes
much more sense to have functions interp1d and interp2d, each with a
variety of interpolation methods available.
> Another problem is for value_locate. Some suggestions
> mentioned to use value_locate. Here is a example to show my problem.
>
> IDL> vec = [2.0, 5.0, 8.0, 10.0]
> IDL> print, vec
> 2.00000 5.00000 8.00000 10.0000
> IDL> loc = VALUE_LOCATE(vec, [0.0, 4.5, 5.0, 6.0, 12. ])
DL> print, loc
-1 0 1 1 3
VALUE_LOCATE is doing what it says it does -- returning a value j such
that
vec[j] < x < vec[j+1]. I don't think anyone suggested that
VALUE_LOCATE can give you the answer by itself, but both JD Smith and
the archive posting from David Fanning (http://tinyurl.com/r9t5s)
showed how you could use VALUE_LOCATE to get
the index of the nearest value. In this case
loc = round(loc+(x-vec[loc]) / (vec[loc+1]-vec[loc]))
(Actually one should first check that 0 < loc < N_Elements(vec)-1, as
in
http://idlastro.gsfc.nasa.gov/ftp/pro/math/tabinv.pro )
--Wayne
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
