[pdal] Bug in readers.gdal?

[Kristian Evers]

Sorry for not getting back to you sooner, Andrew.

I’ve rebuild PDAL so it now has your recent changes to the GDAL reader. It doesn’t look quite right to me. I’ve made a gist with some pipeline xml files and outputs from pdal info: https://gist.github.com/kbevers/22ae84f9ab60e5789587. In short, I’ve set up two pipelines that translate between tif and laz. The first pipeline is just a readers.gdal -> ferry band-1=Z -> writers.laz. The other is readers.gdal -> ferry band-1=Z -> range Z[0.0:100.0] -> writers.laz.

The output laz-files are found on dropbox: https://www.dropbox.com/s/8zyx8kka0mpr7ir/test_gdal_reader_ferry.laz and https://www.dropbox.com/s/fj35lzulu2vvgdc/test_gdal_reader_range.laz

In both files you see that not all Z-values have been read properly by the gdal reader. There is a narrow strip of data in the western part of the file that has the correct Z-values. Moving east all Z-values are zero which they obviously shouldn’t be. You see it best in test_gdal_reader_range.laz but if you inspect the other laz-file carefully you see the same.

/Kristian

Fra: Andrew Bell [mailto:andrew.bell.ia at gmail.com]
Sendt: mandag 08. februar 2016 20:59
Til: Kristian Evers
Emne: Re: [pdal] Bug in readers.gdal?

On Mon, Feb 8, 2016 at 9:20 AM, Kristian Evers <kreve at sdfe.dk<mailto:kreve at sdfe.dk>> wrote:
Okay. I haven’t looked into the code but my feeling is that it is more than just an error in the documentation. If the problem was only the docs I would expect the pipeline to work when switching the in/out dimensions. Would you agree?

I found a couple of bugs that were causing trouble with your file.  I pushed a fix to the PDAL master branch.  Please note that your file contains NODATA values.  PDAL does not currently support NODATA value, but you can use a PDAL filter to remove points that contain NODATA values if you like.

Best,

--
Andrew Bell
andrew.bell.ia at gmail.com<mailto:andrew.bell.ia at gmail.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20160210/2bb37246/attachment.html>