| Re: Reproject MODIS automatically using IDL [message #45355] |
Tue, 06 September 2005 08:07 |
James Kuyper
Messages: 425
Registered: March 2000
|
Senior Member |
|
|
zamri79@gmail.com wrote:
> I looking at Modis 1B. I want to do a reproject to WGS84 (Malaysia
> Datum).
I don't know anything about "Malaysia Datum". However, the keyword
DATUM=8 causes MAP_PROJ_INIT to use the WGS84 Datum for map
projections. If you're using MAP_SET, it's a little trickier, because
you have to use the ELLIPSOID keyword, and you have to know the actual
numbers that correspond to the WGS84 Datum, and the ELLIPSOID keyword
only affects the Transverse Mercator and Lambert Conic projections. The
following IDL code is supposed to set up suitable contents for a WGS84
ellipsoid, but it hasn't been independently verified:
f = 1-6356752.300000/6378137.000000
eccentricity = 2*f-f2
ellipsoid = [6378137.000000, eccentricity, 0.999600]
|
|
|
|
| Re: Reproject MODIS automatically using IDL [message #45359 is a reply to message #45355] |
Mon, 05 September 2005 22:56  |
peter.albert@gmx.de
Messages: 108
Registered: July 2005
|
Senior Member |
|
|
If you have to do serious reprojection tasks, it might be worth having
a close look at proj (http://www.remotesensing.org/proj/). It has
nothing to do with IDL, though, but is probably the tool you are
looking for. As far as I know, IDL only supports WGS84 datum, and if
you have to do datum shifts, you can't do it with IDL. However, I am
also curious to know what the difference is between "classical" WGS84
and "Malaysia Datum" ...
Regards,
Peter
|
|
|
|
| Re: Reproject MODIS automatically using IDL [message #45362 is a reply to message #45359] |
Mon, 05 September 2005 17:54 |
James Kuyper
Messages: 425
Registered: March 2000
|
Senior Member |
|
|
zamri79@gmail.com wrote:
> I looking at Modis 1B. I want to do a reproject to WGS84 (Malaysia
> Datum).
> So i need to this everyday automatically. I'm new in IDL and and i need
> you all to assist me or give some instruction or maybe some coding to
> do that.
I'm familiar with WGS84 as a spheroid, and the geolocation that was
copied into the MODIS L1B was calculated using the WGS84 spheroid. Is
WGS84 also a map projection? I'm not familiar with that projection.
|
|
|
|
|
|
|
|
| Re: Reproject MODIS automatically using IDL [message #45376 is a reply to message #45368] |
Fri, 02 September 2005 14:08 |
George N. White III
Messages: 56
Registered: September 2000
|
Member |
|
|
On Fri, 2 Sep 2005, James Kuyper wrote:
> zamri79@gmail.com wrote:
>> Hallo,
>>
>> Can someone give some instruction or maybe some function/code to
>> reproject MODIS data automatically?
>
> What kind of MODIS data are you looking at?
Excellent question -- so far I've encountered at 3 types:
1) MODAP? data from standard land/atmosphere processing also early ocean
data
2) MODIS ocean data processed with NASA's SeaDAS (much of which is written
in IDL)
3) direct broadcast
When someone comes into your office holding a CD labelled "MODIS DATA" how
do you know what they have?
--
George N. White III <aa056@chebucto.ns.ca>
|
|
|
|
| Re: Reproject MODIS automatically using IDL [message #45377 is a reply to message #45376] |
Fri, 02 September 2005 10:06 |
James Kuyper
Messages: 425
Registered: March 2000
|
Senior Member |
|
|
zamri79@gmail.com wrote:
> Hallo,
>
> Can someone give some instruction or maybe some function/code to
> reproject MODIS data automatically?
What kind of MODIS data are you looking at?
The Level 2 products are arranged in 5-minute granules, and have the
same organization as the raw images that the were developed from: 203
or 204 scans per granule, 1354 1km frames per scan, 10 detectors/frame
of data collected for each 1km band, 20 detectors by 2 samples/frame
collected for each 500m band, and 40 detectors by 4 samples/frame
collected for each 250m band.
For best results with level 2 products, be sure to use the Geolocation
data at 1km resolution that is present in the MOD03 (Terra) or MYD03
(Aqua) files. Keep in mind that each scan should be treated as a
seperate image: consecutive scans overlap each other by 0% at the
middle of the scan, increasing steadily with distance from the center
to 100% at each end of the scan. It makes absolutely no sense to
interpolate anything across scan boundaries. If you need any resolution
higher than 5km, you should NOT interpolate the 5km geolocation sub-set
that is present in the MOD021KM and MYD021KM files; the interpolation
will inherently mis-represent the overlap of consecutive scans.
The MODIS products at level 3 and higher are generally gridded, using
the HDF-EOS Grid constructs. IDL provides wrappers for the HDF-EOS
routines that will provide you with all the information you need to
determine the parameters of the map projection on which the grid points
appear as a uniformly spaced rectangular grid. It's fairly straight
forward, to use EOS_GD_INTERPOLATE to interpolate the data to an
arbitrary set of positions, and in particular to interpolate it to grid
points for a different map projection.
As the person responsible for MODIS geolocation, I've had surprisingly
little need to actually display images. The only code I could give you
is hopelessly specific to the problem I was trying to solve. It was
also my first (and so far, only) attempt at a sophisticated widget
program, and it stopped working shortly after I stopped needing it,
when we upgraded our version of IDL. I don't think it's really a good
example for you to copy. Luckily, there are other people on this
newsgroup who almost certainly will be responding soon with more
specific help on this issue.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
