[gdal-dev] Black borders after reprojection GDAL_WARP

[Even Rouault]

Black borders are completely expected: reprojection changes the geometry 
of the image.

That said, if the area of interest is sufficiently small and the 
geometry change being approximatively an affine transformation, you can 
try https://gdal.org/programs/gdalmove.html instead of gdalwarp to only 
alter the georeferencing information and not touch pixel values at all.

Even

Le 18/11/2021 à 15:34, Lorenzo Di Giacomo a écrit :
> Hi Carl, thanks for your reply, i noticed that it happens even if i 
> just reproject the image, without cutting.
> Of course the dstalpha works, but it increases the size of the image 
> and it changes its nature (adding another band) since this operation 
> is an intermediate operation the resulting image can't be different 
> from the original, just reporojected.
>
>
>
> Il giorno gio 18 nov 2021 alle ore 14:11 Carl Godkin 
> <cgodkin at gmail.com <mailto:cgodkin at gmail.com>> ha scritto:
>
>      Hi Lorenzo,
>
>     I have faced this and there are two things to consider.
>
>     First, you can specify the georeferenced extents _and_ the SRS  of
>     the extents using something like this:
>
>         gdalwarp -t_srs EPSG:3857 \
>             -te_srs EPSG:4326 -te -109 32 -102 36 \
>             input.tif output.tif
>
>     Note that I'm warping to one coordinate system but trimming based
>     on another one.  Is it possible that your black boundary is due to
>     trimming in the wrong coordinate system?  For instance, if your
>     input map is projected but the boundary of the map consists of
>     parallels of latitude or meridians of longitude, then you could
>     use something like the above.  (You can actually get even fancier
>     by trimming with polygons too; see this example for inspiration:
>     https://gis.stackexchange.com/questions/45053/gdalwarp-cutline-along-with-shapefile
>     <https://gis.stackexchange.com/questions/45053/gdalwarp-cutline-along-with-shapefile>
>     .)
>
>     Second, recognize that the output will always be a rectangle in
>     the output SRS.  If the black border is due to "no data" areas
>     outside the input map's extent appearing in the output map, then
>     you can use -dstalpha ("Create an output alpha band to identify
>     nodata (unset/transparent) pixels.") to mark the output pixels
>     that shouldn't be part of the map.  Basically your black pixels
>     become transparent in this case.
>
>     I hope that helps,
>     carl
>
>
>     On Wed, Nov 17, 2021 at 8:11 AM Lorenzo Di Giacomo
>     <loridigia at gmail.com <mailto:loridigia at gmail.com>> wrote:
>
>         Hi all, i have a question, how is it possible to avoid black
>         borders after reprojection with "gdal_warp" ?
>         Basically i have an image that is little rotated, when i
>         reprojected from 32632 to 4326 the resulting image has black
>         borders more or less tight depending on the rotation.
>         I saw i can change the colors of those no_data (using
>         -dstnodata) but how can i do if i dont want it at all? Adding
>         another band result in a size increase, that i don't want either.
>
>         Thanks !!
>         _______________________________________________
>         gdal-dev mailing list
>         gdal-dev at lists.osgeo.org <mailto:gdal-dev at lists.osgeo.org>
>         https://lists.osgeo.org/mailman/listinfo/gdal-dev
>         <https://lists.osgeo.org/mailman/listinfo/gdal-dev>
>
>     _______________________________________________
>     gdal-dev mailing list
>     gdal-dev at lists.osgeo.org <mailto:gdal-dev at lists.osgeo.org>
>     https://lists.osgeo.org/mailman/listinfo/gdal-dev
>     <https://lists.osgeo.org/mailman/listinfo/gdal-dev>
>
>
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/gdal-dev

-- 
http://www.spatialys.com
My software is free, but my time generally not.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20211118/57c5c57e/attachment-0001.html>