Open NasH5169 opened 2 years ago
Hi @NasH5169, I've faced a very similar case, and it was caused by the nodata values. As you do, I gave them a None value, but when I left the nodata value that the data already had I didn't found any rare artifacts. Perhaps it could help you.
Same problem here. Checking the rgb values in photoshop, it seems the calculation is broken.
Hi,
We are trying to display our custom terrain / hillshade using RGB tiles. To do so, we went through the following procedure:
gdalwarp -t_srs EPSG:3857 -dstnodata None -novshiftgrid -co TILED=YES -co COMPRESS=LZW -co BIGTIFF=IF_NEEDED -ts 18000 18000 -r lanczos -wo SOURCE_EXTRA=1000 SRTM.tif SRTM.3857.tif
rio rgbify -b -10000 -i 0.1 SRTM.3857.tif SRTM.3857.RGB.tif
gdal2tiles.py --zoom=0-12 --processes=8 --srcnodata=-32768 -x SRTM.3857.RGB.tif tiles/
mb-util --image_format=png --scheme=tms tiles/ terrain_rgb.mbtiles
So, we ended up with a working MBTiles filled with RGB PNG files, though below the result is not satisfying as you can see different weird spots usually located at around 3000 meters height.
We tried the procedure with two different sources (SRTM and EUDEM) the result is the same.
Do you have any idea?
Thanks.