Cloud to mesh distances

Feel free to ask any question here
Post Reply
Isa
Posts: 1
Joined: Thu Aug 10, 2017 2:22 pm

Cloud to mesh distances

Post by Isa »

Hi all,

I am currently using cloud to mesh distances in my project and I have some doubts:

1. How is the cloud to mesh distances calculated?. I read the wiki page where it says that "for each point of the compared cloud,
CloudCompare will simply search the nearest trinagle in the reference mesh" (http://www.cloudcompare.org/doc/wiki/in ... omputation). But how is the distance exactly calculated? Is it similar to cloud to cloud distances with local modeling? I found Daniel Girardeau-Montaut's thesis, where it is explained in detailed in Section 2.2 but my knowledge of French is almost 0. Is there any paper about cloud to mesh distances?

2. The mean value obtained after the cloud to mesh or cloud to cloud distances calculation is the RMS value?. At the beginning I thought it was the mean of a vector conformed by the distances calculated as d(p,S') = min ||p-p'|| (see eq 1 in Change detection on point cloud data acquired with a ground laser scanner), but then I found this thread http://www.danielgm.net/cc/forum/viewtopic.php?t=1296, where Daniel says: "You can compute a full RMS afterwards by computing the distance between both clouds (with Tools > Distances > Cloud/Cloud dist.')." and I got confused.

Thanks in advance,

Isabel
daniel
Site Admin
Posts: 7332
Joined: Wed Oct 13, 2010 7:34 am
Location: Grenoble, France
Contact:

Re: Cloud to mesh distances

Post by daniel »

Hi,

1. The distance to the nearest triangle is defined as either the orthogonal distance from the point to the triangle plane, if the orthogonal projection of the point on this plane falls inside the triangle, otherwise it's the distance to the nearest edge that is taken.

I believe it's described in the 'Metro' algorithm by Cignoni et al., and you can also find a description here: https://www.geometrictools.com/Document ... angle3.pdf

2. No, the mean distance is indeed the mean of the vector of all distances. As well as the standard deviation. But you can deduce the RMS from these two values. And in the latest versions (2.9 at least), if you launch the 'Edit > SF > Compute Stat. params' on the cloud with the distances, it will automatically output in the Console the RMS of the active scalar field (i.e. the distances in this case).
Daniel, CloudCompare admin
Post Reply