> Date: Wed, 3 Jan 2018 04:42:50 -0600 (CST)
> From: franco_bez <[hidden email]>
> To: [hidden email] > Subject: Re: [mkgmap-dev] DEM-file generation from geo-tiff sources ?
> Frank Stinner wrote
>> At the moment we have to translate the tiffs from EPSG:3035 to EPSG:4326,
>> then the segmentation in 1?x1? tiffs an at last the conversion to hgt
>> files. This should not be a very great problem with gdal-tools or so, but
>> it is additional work.
> Hi Frank,
> thanks for the hint
> It was a long process (some hours), but in the end I succeeded in converting
> 4 of the giant copernicus tiff-tiles to 627 small (1?x1?) hgt files.
> 1. I joined the 4 tiles to one monster-file using gdal_merge
> 2. I converted the coordiante system with "gdalwarp -s_srs EPSG:3035 -t_srs
> 3. with a small shell script the file got cut to 1?x1? tiles using
> i.e. "gdal_translate -strict -q -eco -projwin 10 48 11 47 " for the
> N47E010 tile
> 4. with another script the tiles were converted from tif to hgt format also
> using gdal_translate
> "gdal_translate -strict -q -eco -of SRTMHGT -outsize 3601 3601 "
> the scripts I used are only "quick 'n' dirty" hacks.
> a few notes:
> There were a few problems to work around
> the coordinate system in use by copernicus does not follow the
> latitude/longitude degrees
> Corner Coordinates of the eu_dem_v11_E30N30.TIF:
> Upper Left ( 3000000.000, 4000000.000) ( 12?15'58.58"W, 57?13'49.19"N)
> Lower Left ( 3000000.000, 3000000.000) ( 8? 7'39.52"W, 48?38'23.47"N)
> Upper Right ( 4000000.000, 4000000.000) ( 4?25'17.11"E, 58?59'15.42"N)
> Lower Right ( 4000000.000, 3000000.000) ( 5?31' 4.07"E, 50? 1'26.83"N)
> in EPSG:3035 coordinate it's a perfect square, but in EPSG:4326 it's a
> I decided to join the copernicus tiles before converting the coordinate
> system in order not to run into any problems at the original tile borders.
> the hgt files must start precisely at the degree borders, so the tiles need
> to be cut at the precise "projwin" coordinates
> last the hgt tiles must be 3601x3601 pixel
I've been following this thread and I want to send congratulations for all the hard work to everyone who has made this DEM feature possible. I look forward to giving it a try when Gerd finishes the documentation.