[pdal] PDAL pipeline issues with large datasets

[Howard Butler]

> On Aug 5, 2022, at 3:33 PM, Nandigam, Viswanath <vnandigam at ucsd.edu> wrote:
> 
> Hi All,
> I’m trying to get pdal pipeline working with the latest install of pdal via conda. It works for very small areas but is erroring out on larger selections.
> 
> The pipeline.json file is
> 
> {
>   "pipeline": [{
>     "type": "readers.ept",
>     "filename": "https://s3-us-west-2.amazonaws.com/usgs-lidar-public/USGS_LPC_TN_Middle_B2_2018_LAS_2019/ept.json",
>     "bounds": "([-9666713.882998705, -9662911.140841518], [4283077.817410117,4287281.853965801])"
>   },
>   {
>     "type":"filters.reprojection",
>     "in_srs":"EPSG:3857",
>     "out_srs":"EPSG:32616"
>   },{
>     "type" : "writers.las",
>     "filename": "points.laz",
>     "a_srs": "EPSG:32616"
>   }]}
> 
> 
> The version of pdal (via conda) is
> 
> pdal --version
> -----------------------------------------------------
> pdal 2.4.2 (git-version: e4772e)
> -----------------------------------------------------
> 
> I have tried it on two machines with both Ubuntu 22.04 LTS and Ubuntu 18.04.
> It seems to work with PDAL version 2.4.1.
> 
> I see the following error:
>> PDAL: readers.las: VLR 1(SF/0) size too large -- flows into point data.
> 
> Maybe I’m missing something. Is anyone else having this issue? Can anyone please advice? 
> Thank you!

I think the issue is one of the LAZ tiles in that EPT is bad.