CGAL users discussion list

[Simon Giraudot <>]

Hello,

In the current implementation of WLOP in CGAL, a new set of points is 
generated that is not directly relatable to the input point set, so 
there is no direct way of propagating color information (or texture).

My advise would be to do some postprocessing. You can use a KD tree 
containing your input point cloud and use each WLOP output point as a 
query to search for closest points: in that case, you can relate an 
uncolored WLOP output point to one (or several) input points and 
propagate (or compute average) colors. You can, for example, use a 
counting iterator as input to still get access to the point's color from 
the KD tree, see: 
https://doc.cgal.org/latest/Spatial_searching/index.html#title12

Best,

--
Simon Giraudot, PhD
R&D Engineer
GeometryFactory - http://geometryfactory.com/


Le 11/02/2018 à 10:04, williamlai3a a écrit :
> Hi All,
>
> I have been trying out the function wlop_simplify_and_regularize_point_set()
> in CGAL v4.12 master build, trying to simplify and regularize my very dense
> scanned point cloud in .las format, where it comes with RGBI property.
> The result looks very good, however, I am facing two issues unsolved, and
> would like to seek advise:
>
> 1. The output of wlop_simplify_and_regularize_point_set() is only a list of
> points with only position information. From my understanding on how WLOP
> works, my points are not just sampled, but some of them also re-positioned.
> Thus, how am I suppose to map back my RGBI from my original point set to the
> new WLOP processed point set? Or otherwise, the color info is lost.
>
> 2. Instead of color, I might later on want to input a point set with texture
> coordinates. Similar with problem 1 above, how am I suppose to retain (or
> remap) the texture coordinates to the new WLOP processed point set?
>
> Thank you, and hope I could get some advise here.
>
> William
>
>
>
> --
> Sent from: http://cgal-discuss.949826.n4.nabble.com/