Distance Computation: Approx results?

[whistleboy]

Greetings,

First, THANK YOU for the amazing software! Exactly what I've been looking for to validate some research methods. Here is my situation. I am comparing two point clouds with ~89 points. The point clouds are based on the positions of markers on an object that were captured with two different methods (photogrammetry v. survey instruments). Therefore, the points in one cloud should line up with the points in the second cloud. I want to determine how well the two methods agree in identifying the locations of the point. I have been able to open in the two clouds, synchronize, align and compute Cloud/Cloud Distance. However, I have a few questions that I've not found answers to when looking through the documentation or in the forums:

QUESTIONS:
1) In "Approx. results" what does "Max dist." represent? The value seems much larger than any of the individual point/point distances between clouds.

2) If I'm reporting differences between corresponding point locations in the clouds, will any of the numbers in "Approx. results" be important to me, or should I instead focus on the numbers in "Precise Results"?

Thank you!

[daniel]

Hi,

in fact, approximated distances are only computed by CC to set automatically the precise distances computation parameters (mainly the octree level). It is suited for very large clouds (millions of points) with rather important differences (for those clouds, this parameter is very important as it can drastically lower - or increase - the computation time). It is also a very coarse approximation, especially for very close clouds. Most of the users should ignore them (and this is definitely your case).

You should indeed concentrate on the precise distances. Once they are computed, you can close the dialog and use the various tools to 'look' at them (color scale, point picking etc.) or export the corresponding scalar field to a file (most probably the ASCII format, in which the distances are saved along points - one point per line: 'X Y Z distance').

Best

[whistleboy]

Again, thank you so much for the quick reply and great software. We will be submitting a manuscript in a few weeks, referenceing CloudCompare, and will make sure to forward you a copy when it is accepted for publication. :-)