[pdal] GDAL Writer artifacts along edge of tile

[Andrew Bell]

Hi,

I'm not sure that there's enough information here for me to help you.
You'll have to be specific about how the counts are different than you
expect, how you're creating your tiles and reassembling them.  Keep in mind
that points outside the provided bounds aren't considered when making the
raster.  This typically means that some points that would be considered as
contributing to a cell will get discarded when that cell is on the edge of
the raster.  You're probably better off making the bounds one cell larger
than you need (depending on the radius argument), and then trimming those
edge cells away with GDAL or some such.  This is behavior that could be
changed, but I believe that when I wrote the code, I was trying to keep it
consistent with points2grid, which was the software that PDAL originally
used to create rasters.

Hope that helps,


On Mon, Aug 13, 2018 at 8:15 AM, Lars <laasunde at hotmail.com> wrote:

> Hello,
>
>
> We have created a pipeline using pgpointcloud reader and gdal writer (see
> attached file).
>
>
> Using the pipeline to create a full raster for an area works great.
> However if we tile the same area we are seeing some 'artifacts' along the
> edge of the raster (band 5, Count). The count is different along the edges
> of the tiled raster compared with a full raster (see attached screenshot
> taken from QGis showing band 5 of rasters).
>
>
> We use the 'where' option in the reader.pgpointcloud to provide points
> that is writer.gdal.bounds + margin.
>
>
> We would like to know why these 'artifacts' occur and how to prevent them
> from happening?
>
>
> kind regards, Lars
>
>
> _______________________________________________
> pdal mailing list
> pdal at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/pdal
>



-- 
Andrew Bell
andrew.bell.ia at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20180813/c727e3eb/attachment.html>