CloudCompare forum

Hi,

My name is Arturo Delgado and I'm a Graduate student at Texas A & M University. I am trying to conduct marsh edge erosion using TLS data and I think that using the M3C2 algorithm would be best suited for my research than using a cloud to cloud distance comparison. I have gone through and cleaned up the my three points clouds to reflect my area of interest. I am having difficulty understanding what parameters to use and what they actually mean. Can you clarify what a normal is and how that relates to picking which calculation mode to choose (i.e. multi-scale, vertical or horizontal)? What effect does selecting the core points for normal calculation play on the output? How does the orientation play a role in the output? Any help would be greatly appreciated.
Arturo Delgado

HI Arturo,

we're also using M3C2 to study marsh dynamics (erosion/accretion) and it works really well (once you've removed the vegetation).

Have you read the original scientific paper (Lague et al., 2013) ? I strongly recommend to read it to really understand what's the difference between a cloud 2 cloud comparison and M3C2.

Just to better understand : is your marsh is eroding horizontally (as in bank erosion ?) or horizontally, or both ?
This will have some importance as to your choice of normal calculation mode. For instance, if you're interested in the horizontal marsh retreat, you should use the horizontal normal mode. I'd suggest keeping a constant normal scale at first (i.e. not multiscale) to get acquainted with the results. You could use the multi-scale analysis later on.

Computing the normal with the core points is just an option to speed up the calculation in case your using really dense scans (e.g. cm resolution) and your normal scale is more typically around 2-3 m.

The orientation can play a role in the output: it depends how complex is you marsh boundary.

If you can post a screen capture of your data, I'd be able to help you with setting up M3C2 for your application.

Dimitri

Dimitri,

I am currently reading the scientific paper by Lague et al. to get a better understanding of how the algorithms work but its a little intense since I do not have a coding background but I am pushing through it. It does make some sense to me but I just have to absorb it some more to fully understand the material.

The study site is eroding horizontally (bank erosion). I am attaching a screen shot of our February scan data. I look forward to your suggestions on how to analyze the data.

Arturo Delgado