Detect areas of "crystal/pyrite forest" in LiDAR DTM

1396
4
09-12-2013 03:56 PM
ChrisSnyder
Regular Contributor III
Looking for some creative ideas...

Here in Western WA State, we have a lot of dense forest/brush. Often the LiDAR laser has a hard time getting good ground returns in dense shrubberies - which results in a funny looking "crystal forest" shapes in the DTMs, where the facets of the ground surface TIN are readily visible due to low mass ground point densities. Here's a picture:[ATTACH=CONFIG]27414[/ATTACH].

I am trying to come up with a reasonably good/fast method of locating "crystal forest" areas like this, given ONLY the bare earth DEM (not the original LAS/ascii raw point files). The best method I have come up with is to create an aspect raster, and do a 3-5 cell radius focal variety - thus counting the "diversity" of aspect angles in a relatively small analysis window. The idea being that areas of low aspect variety (and a slope >= 5% of so) are indicative of these rather planar TIN facet shapes. One issue is that there are also areas of quite planar hill slopes where the aspect is also very similar (think a large planar hill slope facing north). So, while the aspect focal variety method does a pretty good job finding the crystal forest areas, it also is pretty good at finding the planar hillslopes as well. I already have a mask screeing out flat areas such as lakes/ag field/etc.  so that's been taken care of.

Super bonus points if you have any good ideas!
0 Kudos
4 Replies
curtvprice
MVP Esteemed Contributor
Just a thought: check out the Curvature tool.

For more on Curvature, read the help, and  this ArcGIS Blog post. Curvature measures the 2nd derivative of the surface, that is, the change in slope. Your "crystal topography" edges I would expect to have absolute values of planiform curvature, since the slope changes very quickly around those edges.

Maybe you could find those "edge cells" and filter them as another way to prospect for crystal forests.
0 Kudos
ChrisSnyder
Regular Contributor III
Thanks Curtis - that is a good idea. I should add that (unfortunatly) these elevation rasters are in integer feet, which tends to cause some "stair stepping" in the output derivative rasters (aspect, plan/prof curvature, hillshade, etc.). So for example, even the aspect pixels of one of the facets of the "crystal" does not entirely end up being a uniform value like you might expect. Had these been float rasters some of these detection methods would be quite a bit more discriminating I think. Best I have come up with is to add some additional masks - so for example, make sure it is shrubby/forested landcover & a low aspect value diversity & not a flat slope & relatively planar hillslope. Getting a bit better, but not a lot of hope for this one.
0 Kudos
JimCousins
MVP Regular Contributor
What about 3D Analyst Tools>>Conversion>>From TIN>>TIN to Triangle. This creates polygons of the faces of the TIN, and adds an area field. The "Crystal Forest" has much larger face area than the rest of the dataset, so sort area descending, and select the ones above a reasonable cut off value, and these should be your problem areas.
Regards,
Jim
0 Kudos
ChrisSnyder
Regular Contributor III
Jim - Per the TIN idea, yes that would work great, but alas I only have access to the elevation rasters, and not the original pnts that were used... Otherwise it'd be very easy to do what I want! (which again is to identify areas that have a low desity of ground returns - using only a bare earth elevation raster as input).
0 Kudos