Project

General

Profile

SimpleTree output questions

Added by Hackenberg Jan over 3 years ago

Dear users who are interested in using my plugin. I received an email which I prefer to answer here, because other users might have similar questions regarding the output:

1) I really like the .ply folder “good” and I would like to work with this cylinders, but i have problems to identify them in the csv.
If i remember propertly in one of your tutorials you said that these cylinders are the one that have spherefollowing detection, not attractor detection (in detailed.csv) but how about the taper?

Yes I said this. Good catch. In general you can rely on every cylinder in the allom improved models only when
a) You scan pure hardwoood plots in leave of conditions and have a good scanning cover to assure not much of your target tree clouds is occluded. I work of course also with trees having leaves / needles. Depending on the plot you can have still good cylinders also within branch order 1 or 2, but in that case I recommend purely focusing on the stem. The worse the scans are (noise + occlusion in the cloud cause this), the more errors are contained here. Cylinders tagged with attractor are never I direct geometrical fit. Only the spherefollowing cylinders are. That is why you should remove them. Before the allometric correction is applied, I also apply a linear taper function to the stem. If I can detect outliers by this function on the stem I correct them by simply grabing the taper radius and replacing the fitted radius.

Linear tapers are not the best tapers. Within Computree I do not have the statistical tools I have in R and that is the only tapers I can use here.

So:
If you only want good measurements derived by cylinder fitting, you need to remove also cylinders tagged with “TAPER”. In that case I would also recommend you to not use my allometric correction step at all. The routine improves the twigs, but the stem has only 1 tip and that is ignoreable. So applying allometric correction should not be needed here. Only keep spherefollowing cylinders from the first modelling. Build your own taper, remove outliers from that function, refit the tatper and get stem volume from your refitted taper. You need to be more experienced to do so in R.

2) what are the differences between DBH and DBH_taper_linear in the result_list.csv? How do you estimate DBH_taper_linea?
Because in some trees the diference is quite important, even in some trees one of this parameter is 0 but not the other.

The DBH field is always filled with a positive value in case there is a fitted stem cylinder in DBH height. Sometimes the pipeline extracts though a tree higher than 1.3 meter, lets say from 2 meter height on. The lower points were lost during one of the filterings. In that case the lowest cylinder is at height 2 and there is no direct DBH cylinder and you will get a 0. But as explained above if you have a stem description ranging from 2-20 meters you can still extrapolate by tapering. I do so, I fit my linear taper and grab the diameter from that function.

So you should always take DBH in case it exists. This is a direct measurement of DBH, not a statistical (taper) one. In case the value is zero, just take DBH_taper_linea. Even if it is not the best taper I can fit within CT, this value is better than zero.

I hope this helps,
Jan