CloudCompare forum

Hello CC community, hello Daniel;

I'm working with scans of pavement surface, my aim is to detect the surface unevenness: I mean the presence of bump and/or depression.
Up to now I used the Roughness tools but, as I learnt, the roughness value is always positive, since CC consider the absolute value to the best fitting plane. Therefore, I can't deduce the correct 'sign' of the distance, distinguishing the fact that it is a bump or a depression.

Now my questions:
1. Is there a way to assess if the point cloud is above or below the best fitting plane?
2. In your opinion, is there a better tool to be used in order to assess if the surface (limited to the Kernel size area) is convex or concave?

I hope I'm clear enough.
Thank for your help

Yes, that's an interesting application.

And there's currently no way to do what you want...

We could add an optional parameter to define the 'vertical' direction. Then, the algorithm would be able to sign the value after making sure the fitted plane is in the right orientation. It would not work in the general case, but for a 2.5D cloud such as a road, it should work I guess.

Here's what the prototype looks like (with a radius of 4):

For information, this is now available in the latest 2.12.beta version!

Here's another example:

Perfect, thanks.
I'll try immediately.