[Gdal-dev] goode homolosine projection

[Vincent Schut]

On Thursday 11 March 2004 08:56, Vincent Schut wrote:
> Manuel, Frank,
>
> I just want to say thank you, this is what I have been waiting for :-)
> Till now I tried reprojecting modis datasets from goode's homolosine by
> cutting the image into pieces which I seperately reprojected using either a
> sinusoidal or a mollweide projection, using the parameters from
> http://gis.esri.com/library/userconf/proc98/PROCEED/TO850/PAP844/P844.HTM.
> This patch will make my life easier...
>
> One question though, because you mention reprojecting modis datasets. I
> read somewhere on the proj mailinglist
> (http://remotesensing.org/lists/proj_archive/msg03772.html) that for
> reprojecting from different ellipses (sphere -> WGS84), one needs to
> specify the '+towgs84=0,0,0' parameter too. Do you know it this is correct
> and still necessary? I think the correct phrase to reproject modis from
> goode's homolosine (sphere) to longlat WGS84 then will be something like:
>
> gdalwarp <input> -s_srs '+proj=goode +ellps=sphere +towgs84=0,0,0
> +lon_0=100w +x_0=-11119487.43'  <output options>
>
> Needless to say I hope this patch will be in cvs soon... But I'll apply the
> patch locally and test it a bit on my modis data of Indonesia, and keep you
> informed on the results (no news is good news is an applicable dutch saying
> here).

Results of some quick testing: Things seem to work correctly when using an 
additional '+datum=sphere +towgs84=0,0,0' for projection modis 32 day 
composites from goode's to latlong. NB when omitting the datum/towgs thing 
there is an erronous shift along the equator, which gets visible when warping 
a scene that is both on the northern and southern hemisphere (in my case: 
Kalimantan, Indonesia), probably due to lack of/bad datum/sphere transform?

Cheers,
Vincent.


-- 
-----------------------------
 Vincent Schut
 SarVision B.V.
 Wageningen, The Netherlands
 www.sarvision.com