Very cool, Even! In the end I managed to come up with a solution that uses +geoidgrids but this will certainly make things easier in the future. Thanks for taking the time to implement this!
/Kristian > -----Original Message----- > From: Even Rouault <[email protected]> > Sent: 11. oktober 2021 23:36 > To: Kristian Evers <[email protected]>; [email protected] > Subject: Re: [gdal-dev] Custom vertical transformations > > Kristian, > > https://github.com/OSGeo/gdal/pull/4639 will allow you to use an > arbitrary pipeline to do a vertical transformation through gdalwarp > > Even > > Le 23/08/2021 à 10:50, Kristian Evers a écrit : > > Thanks for your insights, Even! I had a feeling I was pushing the boundaries > here, good to get that confirmed. > > > > The idea of using -s_srs and a correction grid is intriguing. I assume this > > way > will be easier to implement as well since I don't have to worry as much about > the > CRS and extent of my correction grid as when using gdal_calc. The correction > grids I already have are in EPSG:4326 and I assume GDAL is clever enough to > handle that or at least pass the responsibility over to PROJ. I'll try a few > things > and report back how it goes. > > > > I'll make sure to add a ticket for extending the scope of gdalwarp to the > > 3D- > space. > > > > /Kristian > > > >> -----Original Message----- > >> From: Even Rouault <[email protected]> > >> Sent: 20. august 2021 17:30 > >> To: Kristian Evers <[email protected]>; [email protected] > >> Subject: Re: [gdal-dev] Custom vertical transformations > >> > >> > >>> Regarding your immediate issue, the workaround I see would be that you > >>> use PROJ to create a regular grid that contains the vertical > >>> correction, use gdalwarp to transform it to the exact extent and > >>> resolution of the result of your 2D-only gdalwarp, and finally use > >>> gdal_calc to sum the two. > >>> > >> Well, if you actually create a vertical correction only grid, then you > >> could just use -s_srs "+proj=longlat +ellps=... > >> +geoidgrids=your_tif_or_gtx" (or maybe as output, or fix the sign of > >> the correction in the grid), since the workaround I give above is what > >> gdalwarp would do. But of course none of this will solve the > >> time-dependent aspect you need. > >> > >> -- > >> > >> http://www.spatialys.com > >> My software is free, but my time generally not. > > -- > http://www.spatialys.com > My software is free, but my time generally not. _______________________________________________ gdal-dev mailing list [email protected] https://lists.osgeo.org/mailman/listinfo/gdal-dev
