[pgpointcloud] Overlapping patches

[Howard Butler]

On Jan 31, 2014, at 4:22 AM, Nilesh Vyavahare <Nilesh.Vyavahare at SeaZone.com> wrote:

> The amount of overlap I am getting between patches is usually more than 90%. The reason I am more after non-overlapping patches is because I am thinking of first comparing patch extents with a geometry to work out intersecting points, which hopefully should reduce time to work out intersecting points with a geometry. Or do you think it takes about the same time to work out the intersection either you compare patch extent or patch itself?

You should not be getting *any* overlap for single-run chipper calls. Can you describe more completely how you are generating your inserts and coming  up with a 90% overlap value?


> Thanks for this Paul. I am probably being bit more demanding but the extent given by this way is an MBR, can I get hexbin extent in pgpointcloud similar to the ones shown in Howard’s PDAL hexbin filters?

There's a method called --pipeline-serialization that can be used to get the resultant hexbin of a pipeline invocation. Combine that with some shell scripting and you can insert the geometry into your database *after* your pipeline has run.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 495 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.osgeo.org/pipermail/pgpointcloud/attachments/20140131/3bbc3517/attachment-0001.pgp>