AW: [gdal-dev] An issue about using GDALWARP todotheNTv2transformation

[Schmitz, Uwe]

Heng,
> 
> 
> Hi Uwe,
> 
> Thanks for your quick response as usual :) 
> 
> Yes, I had to use the "BIGTIFF=YES" option on some of these 
> large data to get the re-projection done, but did not need to 
> use this option on all the data which had this problem. The 
> word "large" was actually mentioned due to the fact that all 
> the data which do not have this "gap" problem after the 
> re-projection are smaller than the ones which have. If you 
> want, I can give you more details about the actual size of 
> those data tomorrow as I do not have access to the data at the moment.
> 
> I actually can clearly see those "gaps" in FW OpenEV and in 
> one of our own tools (only for viewing). The gaps are always 
> there no matter I zoom in/out them in whichever way. Of 
> course, the gaps will become bigger if I zoom in the picture.
> 
> The original data are actually provided in smaller tiles 
> (GeoTiff) and all I need to do this time is re-projecting the 
> data and displaying them as a whole dataset. However, after I 
> did the re-projection on those tiles separately, there appear 
> to be black lines around those tiles when displaying them in 
> one view/window (even after merging them with GDALWARP 
> command. Shall I use the option "skip_nosource=yes" to mosaic 
> the tiles in the GDALWARP command?). And this was why I chose 
> to merge the smaller tiles first and then re-project the 
> merged large dataset. 
> 
if these gaps are one or two pixel in size and around the
tile boundaries, it is possible that they are caused by 
rounding problems during transformation. I don't know
anything about your data, but If they are RGB try to use 
another interpolation algorithm like "-r cubic". The
command you posted uses "-r near". This can cause those black 
pixels.

Uwe