Thanks Javier for your reply. I'd like to avoid changing the source file work around. We are using gdalwarp C++ API as part of a general application that can open any kind of file.
On Fri, Jul 26, 2024 at 12:26 PM Javier Jimenez Shaw <j...@jimenezshaw.com> wrote: > Hi > > I do not know if there is anything wrong in GDAL. It looks like > > As a workaround, you can first convert the projection to be centred at > lon=0 with something like this (I kept the same radius as the original to > minimize distortion). You can add bilinear interpolation or not, as you > wish. > gdalwarp -t_srs '+proj=eqc +a=2439400 +b=2439400 +lon_0=0.000000 > +lat_0=0.000000 +x_0=0 +y_0=0 +units=m +no_defs'-overwrite > msgr_dem_low_res.tif centred.tif > And then cut with your command. > > I know it is not optimal because it is converting the full image. But the > output is not blur. > > Cheers, > Javier. > > > > On Fri, 26 Jul 2024 at 09:49, Calogero Mauceri via gdal-dev < > gdal-dev@lists.osgeo.org> wrote: > >> Hi, >> >> I'm rendering a 256x256 tile from a mercury dem. I noticed the gdalwarp >> is producing a bad output when using interpolation with the following >> command >> >> gdalwarp -of PNG -r bilinear -t_srs '+proj=eqc +a=2439700 +b=2439700 >> +lon_0=0.000000 +lat_0=0.000000 +x_0=0 +y_0=0 +units=m +no_defs' -ts 256 >> 256 -te -958067.949620 -958067.950079 0.000460 0.000000 -overwrite >> msgr_dem_low_res.tif tile.png >> >> The output image is kind of blurred/low resolution >> [image: tile.png] >> >> On the other side, it looks good if nearest interpolation is used >> >> gdalwarp -of PNG -t_srs '+proj=eqc +a=2439700 +b=2439700 +lon_0=0.000000 >> +lat_0=0.000000 +x_0=0 +y_0=0 +units=m +no_defs' -ts 256 256 -te >> -958067.949620 -958067.950079 0.000460 0.000000 -overwrite >> msgr_dem_low_res.tif tile.png >> >> [image: tile.png] >> >> >> The problem seems related to the fact that the right edge of the tile is >> around longitude 0, while the source image is a full globe/360 degrees >> image centered at 180 longitude >> >> If I reduce the tile size to 255x255 for avoiding the 0 longitude at the >> edge, then the interpolated output file looks much better >> >> gdalwarp -of PNG -r bilinear -t_srs '+proj=eqc +a=2439700 +b=2439700 >> +lon_0=0.000000 +lat_0=0.000000 +x_0=0 +y_0=0 +units=m +no_defs' -ts 255 >> 255 -te -958067.949620 -950583.044220 -7484.905400 0.000000 -overwrite >> msgr_dem_low_res.tif tile_255.png >> >> [image: tile_255.png] >> >> Am I doing something wrong or is this a bug in GDAL? >> >> I'm using the mercury dem from here >> https://pdsimage2.wr.usgs.gov/Messenger/MESSDEM_1001/DEM/GLOBAL/IMG/ >> Here is a lower resolution of the same dem for testing purpose >> https:// >> <https://pdsimage2.wr.usgs.gov/Messenger/MESSDEM_1001/DEM/GLOBAL/IMG/> >> files.actgate.com/terrain/mercury/msgr_dem_low_res.tif >> >> Thanks in advance for your help, >> Calogero Mauceri >> >> -- >> Calogero Mauceri >> Software Engineer >> >> Applied Coherent Technology Corporation (ACT) >> www.actgate.com >> _______________________________________________ >> gdal-dev mailing list >> gdal-dev@lists.osgeo.org >> https://lists.osgeo.org/mailman/listinfo/gdal-dev >> > -- Calogero Mauceri Software Engineer Applied Coherent Technology Corporation (ACT) www.actgate.com
_______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org https://lists.osgeo.org/mailman/listinfo/gdal-dev
