Page 1 of 1

Plane Create/Edit

Posted: Thu Apr 26, 2018 10:21 am
by GnillitsBosch
Hey,

first I want to thank you for this extraordinary program, without this I would not be able or have quite some problems, doing my master thesis.

I have a question regarding the Plane Create/Edit tool. For my thesis I have to compare different orientations of sensors to the Coordinate system of a vehicle. I'm doing this by estimating the plane through the point cloud on the surface of the sensor and afterwards comparing the resulting normal vector to the 3 axis of the coordinate system.
Now I have noticed that when I'm creating the planes of the coordinate system to get the normals, because my vehicle is rotated or something like this, it's really difficult to get them orthogonal to each other. Through testing with some uncomplex planes like (1,0,0), (0,1,0) and (0,0,1) I observed that when I'm creating planes with the primitive factory and afterwards edit them to be orthogonal there's always a small error:
Original:
Position (0,0,0)
Normal (0,0,1)
With the edit tool I changed the normal vector to (1,0,0) but this resulted in:
Position (0,0,0)
Normal (1,-4.37114e-8,-4.37114e-8)

I also found a workaround for this problem, at least for me. When I rotate the plane with the 'Apply Transformation' tool I get the correct plane.

I know this is quite a minor problem, but it's more like that I wanted to know why this happens than that. At least my assumption is that this happens because of computational limits and not because there is a bug or something like this.

Thanks again for this awesome software
Gnillits

Re: Plane Create/Edit

Posted: Thu Apr 26, 2018 7:46 pm
by daniel
Well, this indeed probably due to some numerical accuracy issues computation (which one is not clear to me). However 1.0e-8 is clearly equivalent to 0 from the computer standpoint ;)

I'll still try to detect which unnecessary computation produces this 'round off' leftover.

Re: Plane Create/Edit

Posted: Wed May 02, 2018 7:35 am
by GnillitsBosch
Thanks for the quick response. And also for the confirmation of my theory.

I also know that the 1.0e-8 is negligible especially when my scanner has only an accuracy of 6mm, just wanted to point this issue out to you, and i'm hoping you find the reason for this minor issue :)