[Liblas-devel] Las2Oci problems

[Martin Isenburg]

tom & others,

my quad tree finalization can code could only be used to tile the data
adaptively (with a min or max num of points per quad tree cell).

but with some extra care taken the same code will also be able to allow TIN
streaming sometime later. that would mean that an oracle point cloud
generated with this quad tree code would remain "implicitly finalized" if we
keep some basic info around. both "lastile" and "lassort" use a variation of
this quad tree code as well. currently the quad tree is just used for tiling
and z-order sorting ... but resulting LAS file are also spatially finalized
and could be fed into a streaming pipeline.

http://www.cs.unc.edu/~isenburg/lastools/download/lastile_README.txt
http://www.cs.unc.edu/~isenburg/lastools/download/lassort_README.txt

If it is of interest to (someday) be able to use the point clouds reordered
by libLAS or libPC together with TIN streaming it may be worthwhile to take
that extra care and add a compatible quad tree functionality to the
libraries. Streaming TIN generation relies on overlap tests between circles
and quad-tree cells (which are very very tricky to get right due to
floating-point arithmetic error offsets [1]) so that switching to a
different quad tree code won't be trivial.

cheers,

martin @lastools

[1] Martin Isenburg, Yuanxin Liu, Jonathan Shewchuk, Jack Snoeyink,
Streaming Computation of Delaunay Triangulations, Proceedings of
SIGGRAPH'06, pages 1049-1056, July 2006. http://www.cs.unc.edu/~isenburg/sd/

On Mon, Apr 18, 2011 at 7:18 AM, Michael P. Gerlek <mpg at flaxen.com> wrote:

> > Of course I am willing to provide the source code. Mostly is copied/based
> on
> > SPfinalizer. To whom and where could I send the source to?
>
> Tom:
>
> When you say "copied", do you mean you directly used Martin's source code?
> If so, we cannot accept you contribution, unless Martin's code is also open
> source with an OSI-approved license on it (preferably BSD).  The provenance
> of much of Martin's code is not clear, and libLAS/libPC has to be very
> careful about where we get our code from.  If you only based your code on
> something like a published paper of his, however, that would be okay.
>
> That being said, if your code meets these requirements then please just
> make
> sure your copyright appears on each of the files and you can send it to me
> (mpg at flaxen.com).
>
> Thanks -
>
> -mpg
>
>
> _______________________________________________
> Liblas-devel mailing list
> Liblas-devel at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/liblas-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/liblas-devel/attachments/20110418/d8868b75/attachment.html