[gdal-dev] Discrepancy when tiling on MAC and Linux

[Grégory Bataille]

Hey Even,

Actually a "simpler" solution for now for us is to remove the grid from the
docker image we use to run dataset tiling. This way, the projection follows
TOWGS84.
It's suboptimal in term of absolute accuracy of the tiles, but it gives us
consistency in our data which is more important to us.

Thanks for the support. I don't think I would have found this grid thing on
my own.

---
Gregory Bataille


On Thu, Nov 14, 2019 at 1:03 PM Grégory Bataille <gregory.bataille at gmail.com>
wrote:

> Oh, I could gdalwarp the dataset myself to WGS84 and then tile it, so as
> to consistently "ignore" grids.
> That's an idea indeed.
>
> Thanks for the pointer, I'll take it into consideration
>
> ---
> Gregory Bataille
>
>
> On Thu, Nov 14, 2019 at 12:46 PM Even Rouault <even.rouault at spatialys.com>
> wrote:
>
>> On jeudi 14 novembre 2019 05:54:09 CET Grégory Bataille wrote:
>> > wow, thanks.
>> >
>> > I'm not quite sure I'm following it all though. I'm reaching a bit my
>> > knowledge limit here with those transforms. I'll need to spend a bit
>> more
>> > time on it to get more understanding and see what I can do, how I might
>> use
>> > your update.
>> > But thanks for the quick turnaround
>>
>> Another option if you want full control on the transformation would be to
>> provide the PROJ pipeline directly. gdalwarp has the -ct switch for that,
>> which ultimately uses
>> OGRCoordinateTransformation::SetCoordinateOperation()
>>
>> Even
>>
>> --
>> Spatialys - Geospatial professional services
>> http://www.spatialys.com
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20191114/a63aa780/attachment.html>