[gdal-dev] gdalwarp problems with MODIS L1B

[Even Rouault]

Le jeudi 21 mai 2015 10:19:13, Yuta Sato a écrit :
> Hi,
> 
> I think the MOD021KM swath product is point data, which should be gridded
> using the array of latitude and longitude given.
> gdal griddata may be an option. However, it's very slow even for a single
> band of MOD021KM product.
> 
> MODIS team provides MODIS Swath Reprojection Tool for gridding the data,
> which performs fast.
>  https://lpdaac.usgs.gov/tools/modis_reprojection_tool_swath
> 
> I do not know the science behind MODIS Swath Reprojection Tool why it's
> faster than the gdal griddata.
> Can we discuss and implement a faster gridding tool in gdal? Why I am
> focusing gdal is because the MODIS swath tool works only for MODIS data,
> but the expected gdal tool would work for any data.

Yuta,

Make sure your point layer has a spatial index before running gdal_grid on it. 
You will also need to specify reasonable values for the search radius.

Even

> 
> Thanks.
> 
> On Thu, May 21, 2015 at 4:45 PM, Rutger <kassies at gmail.com> wrote:
> > Hey,
> > 
> > Maybe i'm missing something, but for as far as i know, the MODIS MOD021KM
> > product is still a 'swath', and therefore still containing the bow-tie
> > effect as the result of the scanning sensor. Wouldn't this mean that
> > gdalwarp will never work? The lat/lon arrays aren't continuous, so any
> > fitting (GCP/TPS) wont be accurate. And upsampling the arrays with cubic,
> > like Even did, will result in interpolating pixels which are next each
> > other
> > in the array/file, but aren't next to each other 'on the ground', wont
> > work well.
> > 
> > I think you need to either use a dedicated algorithm (like gradient
> > search),
> > or something very generic (like the KDTree of Pyresample) to overcome the
> > bow-tie effect.
> > 
> > I have spend alot of time experimenting with this in GDAL, since i try to
> > do
> > as much with GDAL as possible. But projecting a swath file from a
> > scanning sensor (like MODIS or VIIRS) is something i have never gotten
> > to work. For some historic discussion about this see for example:
> > 
> > http://osgeo-org.1560.x6.nabble.com/gdal-dev-reproject-python-numpy-binar
> > y-swath-lat-lon-td4978609.html
> > 
> > 
> > Regards,
> > Rutger
> > 
> > 
> > 
> > --
> > View this message in context:
> > http://osgeo-org.1560.x6.nabble.com/gdal-dev-gdalwarp-problems-with-MODIS
> > -L1B-tp5206583p5206825.html Sent from the GDAL - Dev mailing list archive
> > at Nabble.com.
> > _______________________________________________
> > gdal-dev mailing list
> > gdal-dev at lists.osgeo.org
> > http://lists.osgeo.org/mailman/listinfo/gdal-dev

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com