[PATCH] gdalwarp - Polar to Cylindrical reprojections cuts off part of target image when Polar image crosses pole

[glewis]

When using the gdalwarp tool, if converting an image in a Polar projection (such as an laea projection) that crosses the pole, to a Cylindrical projection (such as a latlong projection), by default the target image is cut off and doesn't extend to the pole.

Using -wo settings of SOURCE_EXTRA, SAMPLE_GRID, and SAMPLE_STEP do not change this behavior. Manually setting the cylindrical extent boundaries with -te does work, but since it is manual you have to know what values to use.

Allowing an option to force internal gridding would be beneficial.

[Even Rouault]

Attached a patch that solves the issue.

With it: gdalwarp -t_srs EPSG:4326 polar.gtiff.tif polar_wgs84.tif

gives:

$ gdalinfo polar_wgs84.tif
Driver: GTiff/GeoTIFF
Files: polar_wgs84.tif
Size is 5111, 857
....
Origin = (-180.000000000000000,90.000000000000000)
Pixel Size = (0.070436313832909,-0.070445741950979)
Metadata:
  AREA_OR_POINT=Area
Image Structure Metadata:
  INTERLEAVE=BAND
Corner Coordinates:
Upper Left  (-180.0000000,  90.0000000) (180d 0'0.00"W, 90d 0'0.00"N)
Lower Left  (-180.0000000,  29.6279991) (180d 0'0.00"W, 29d37'40.80"N)
Upper Right ( 180.0000000,  90.0000000) (180d 0'0.00"E, 90d 0'0.00"N)
Lower Right ( 180.0000000,  29.6279991) (180d 0'0.00"E, 29d37'40.80"N)
Center      (  -0.0000000,  59.8139996) (  0d 0'0.00"W, 59d48'50.40"N)
Band 1 Block=5111x1 Type=UInt16, ColorInterp=Gray

There's one part where I'd like to get Frank's opinion. It's the part where I adjust the x and y resolution so that the computed geotransform with the computed width and height doesn't overlap the 180° meridian. The only issue I see with that is that the pixel size is no longer a perfect square, as it's stated in the documentation of GDALSuggestedWarpOutput2(). But otherwise, another patch I'm working on in GDALWarpOperation::ComputeSourceWindow() for dealing with the inverse problem (warping to polar projection) can't work properly if the right edge of the image overlaps the 180° meridian.

[warmerdam]

Remove non-serious milestone.

[Even Rouault]

This ticket has been automatically closed because Trac is no longer used for GDAL bug tracking, since the project has migrated to GitHub. If you believe this ticket is still valid, you may file it to ​https://github.com/OSGeo/gdal/issues if it is not already reported there.