Canopy cover - lascanopy

154 views
Skip to first unread message

Ivan Kljun

unread,
Feb 29, 2016, 6:14:14 PM2/29/16
to last...@googlegroups.com
Hello, 

I am a beginner using LIDAR data for building a habitat model for the nightjar (owl-like bird) in Slovenia:
I am interested in what is the proportion small openings in the forest (vegetation below 3 m). What I would like to get is: (number of sq. meters with vegetation less than 3 m) / the area of forest) which would be a measure of the openness of the forest. The 3 m is because of the ecology of the bird.
Any advice on how to make this?
If I have density and spacing like below, what size can the "step" be? 0,58 or four times this number?
point density: all returns 8.05 last only 2.98 (per square units)
          spacing: all returns 0.35 last only 0.58 (in units)

Best regards,
Ivan Kljun

Antonio Ruiz

unread,
Mar 3, 2016, 7:43:07 AM3/3/16
to last...@googlegroups.com
You get a point on the ground when there is a big enough gap in the canopy (assuming the intensity is also enough). Otherwise you get a return on the vegetation. The penetration index = "ground points" / "total points" is a measure of the proportion of openings. You can extend this definition to other levels, e.g. "points below 5 m" / "total points" to distinguish openings in trees from openings including also bushes. 

Percentiles on height give another description of the vertical distribution of the canopy that could be useful for your application. Pn is the height above ground of an horizontal plane so that n% of the points are above the plane (and (100-n)% are below). P50 is the median of the heights. You could consider you have an opening when the P90 or P95 percentiles are below a certain level (3 m for example). 

You can compute a DTM between 0.5 or 1 m grid step but for the penetration index or the percentiles you need more points per pixel to get some statistical significance. With 4 m grid step you will get almost 129 points/pixel. This would be the minimum acceptable.

I believe you can compute all of this with Fusion but I'm not an expert on Fusion. It is here:


Regards

Irf

unread,
Apr 3, 2016, 10:04:22 PM4/3/16
to LAStools - efficient tools for LiDAR processing
Hello Antonio,

That was nicely explained; however, I have a query regarding "Percentiles on height give another description of the vertical distribution of the canopy that could be useful for your application. Pn is the height above ground of an horizontal plane so that n% of the points are above the plane (and (100-n)% are below).".

In my opinion, Pn is the height above ground of a horizontal plane so that n% of the points are BELOW the plane (and (100-n)% are ABOVE). For example, if the 5th percentile is 9.15 m, this would mean that 5% of the points are BELOW the height 9.15 m. 

I would be happy to have your thoughts on this.

Regards,
Irf
Reply all
Reply to author
Forward
0 new messages