[pdal] Best practice for classifying ground on large cloud?

[michael.smith.erdc at gmail.com]

Another way to do this would be to convert the data to EPT so you can make buffered range requests to the data. You can easily build a geometry tile dataset (qgis has a nice one) and then pass the geometries to a batch pipeline. 

Michael Smith
US Army Corps

> On Aug 12, 2020, at 4:51 PM, Bradley Chambers <brad.chambers at gmail.com> wrote:
> 
> 
> You should also look at 
> https://pdal.io/apps/tile.html if you haven’t already. It helps with buffered tile creation. There will be some work required on your part to delete the buffer and merge the filtered result at the end however. 
> 
> Brad
> 
>> On Wed, Aug 12, 2020 at 14:48 Andrew Cunliffe <andrewmcunliffe at gmail.com> wrote:
>> Does anyone have recommendations for best practice for classifying ground points on large clouds?
>> 
>> I'd like to use SMRF to classify ground points in a large cloud (ca. 920 000 000 points), but am exceeding the system memory (128GB).
>> 
>> It seems logical to split a large cloud into buffered tiles (to minimise edge effects), then run a classifier on each tile, delete the buffer and merge the tiles together again. It seems like this should be possible with filters.splitter, but is this the right way to go?
>> 
>> Any suggestions on implementation much appreciated!
>> Regards,
>> Andy
>> 
>> _______________________________________________
>> pdal mailing list
>> pdal at lists.osgeo.org
>> https://lists.osgeo.org/mailman/listinfo/pdal
> _______________________________________________
> pdal mailing list
> pdal at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/pdal
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20200812/d51050d0/attachment.html>