Cross section slices "bug"

[Luca1988SCENE]

hi, i'm a new member of this Forum. I registered my self here to look for help about using CLUOD COMPARE Software (2.5.3). I'm using this Software for my thesis in Engeneering. I have a cloud of points and i have to extract some profiles. To do this i use the CROSS SECTION tool, i enter my parameters and i get my profiles. I want to understand how the software works to extract theese two profiles and why there are so many straight lines.
I hope I was clear and i will wait for some aim.
i will post some phothos in a second moment.

Thanks a lot.


[img][img]
[/img][/img]

[daniel]

The cross section tool is very simple for the moment. It extracts the "concave hull" of a point cloud (slice) projected in 2D.

Generally you have to play with the "max edge length" parameter (the smaller, the better the polyline will fit the concave hull (even if there's not a unique definition of it).

And indeed, some snapshots would help...

[Luca1988SCENE]

here there are some photos.

Luca

[Luca1988SCENE]

i don' t understand why the profile is splitted in the red circle.

Luca

[daniel]

So you should indeed decrease the "max edge length" parameter to get a more watertight contour.

And the fact is that the tool extracts "closed contours" (i.e. a kind of circular contour that is compressed as close as possible to your points). So if you want to extract a single non-closed line you should say "yes" when CC ask you if you want to split the contour in multiple parts (it might help).

[Luca1988SCENE]

hi Daniel, i want to understand why CC needs the necessity to split the contours. Why the contour doesn t keep on on this figure ?

[daniel]

Obviously the points displayed here are not all used to compute the contour. Can you only display the slice points? And you could also compute the same contour without allowing CC to split it. This way you'll see what the initial contour looks like (before splitting).

[Luca1988SCENE]

Hi Daniel, the creation of profiles is based on a curve fitting process, i would like to know the equation that CC uses to fit this points. if it is possible.
Thanks a lot

[daniel]

Nope, the contours are extracted thanks to a "concave hull" extraction algorithm. For the moment it's a rather simple approach inspired from Andrew's monotone chain 2D convex hull algorithm (retrieved from http://en.wikibooks.org/wiki/Algorithm_ ... tone_chain).