[pdal] Raster to points
Howard Butler
howard at hobu.co
Thu Apr 25 06:16:53 PDT 2019
> On Apr 25, 2019, at 7:29 AM, Steven Spiegel <steven.spiegel at slu.edu> wrote:
>
> Good morning,
>
> I have a question about applying rasters to point clouds. Before I created a TIF, I classified the ground, curvature, height above ground, and normalized intensity. I then exported it as a tif (resolution = 0.1 m, average, etc) without the ground. I then combined the three aforementioned attributes and used some image segmentation techniques to separate the objects in the scene. I now have a labeled raster of my objects (they are not yet classified, just segmented). Is there a way to then apply a label (as opposed to a class) on the points? I know you can colorize with a raster or even if I turned it into a shapefile I can then classify that way, but I'm not quite at that level. I looked through the dimension list (https://pdal.io/dimensions.html <https://pdal.io/dimensions.html>) and didn't see one that would fit for what I'm looking for. Can I create a dimension and then apply the label? If you need further clarification I can provide it. Thanks!
Steven,
filters.colorization is really just a poorly named filters.overlay for raster
https://pdal.io/stages/filters.colorization.html#filters-colorization <https://pdal.io/stages/filters.colorization.html#filters-colorization>
Something like this might work. Use colorization to grab the first band value and stick it in GPSTime (a dimension of type double that should exist in a default LAS schema). Then copy that over to a dimension called "Segmented". Further processing can use the "Segmented" dimension as needed.
> {
> "type":"filters.colorization",
> "dimensions":"GPSTime:1",
> "raster":"segmented.tif"
> },
> {
> "type":"filters.ferry",
> "dimensions":"GPSTime=>Segmented",
> },
Hope this helps,
Howard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20190425/27f0c4e4/attachment.html>
More information about the pdal
mailing list