Page 1 of 1
Command line mode : Subsampling using curvature SF
Posted: Tue Aug 24, 2021 3:42 pm
by ryukisai
Hi there, i’m new here, so first of all Thanks a Lot for the efforts over the years and the great software you wrote!!
I recently discovered that you can use a SF to scale the spatial subsample target of a cloud.
I tried it it the UI, with great results. However, i would like to use it in command line mode to save some time.
Is there already an argument that would allow that?
The wiki page (
https://www.cloudcompare.org/doc/wiki/i ... _line_mode) does not mention that.
Thanks in advance for any suggestion!
Re: Command line mode : Subsampling using curvature SF
Posted: Tue Aug 24, 2021 7:26 pm
by daniel
Ah non, I don't believe this option has been ported to the command line mode! I also believe it's already in the TODO list, but one need to update the code...
Re: Command line mode : Subsampling using curvature SF
Posted: Wed Aug 25, 2021 2:28 pm
by ryukisai
Hi Daniel, thanks for your quick reply!
It’s ok i still very much enjoy this feature in UI mode. It makes a lot of sense to only keep dense data in curvier volumes.
Before that, i used Meshlab’s quadric edge collapse decimation, but having this integrated in CC saves a lot of time and space.
I’ll be looking forward to its CL implementation!
Or maybe i should take a look at the recent Python options at some point…
Merci bien !
Re: Command line mode : Subsampling using curvature SF
Posted: Thu Aug 26, 2021 6:02 pm
by daniel
Ah, good idea!
Re: Command line mode : Subsampling using curvature SF
Posted: Thu Oct 14, 2021 3:53 pm
by ryukisai
Hi there, i’m just quickly reporting after doing many tests on the curvature-dependent SS. It’s a great feature, i love it!
However, since the SS value interpolation is linear, that tends to remove too many points on details (or not enough points on flat areas, depending on your input parameters) for my taste/application.
So i found the trick to perform a logarithm operation on the normal change rate SF, and use the resulting SF as the variable SS input.
I use a max curvature target density of ≈ 1-2 mm and a min curvature target of ≈ 10-20 mm.
It works really well for me so i thought it might be a good tip to share.
Nevertheless, i would love to get a non linear interpolation option directly from the SS window :-D
Alternatively, i see two options to make it better while still keeping linear interpolation:
- entering a single target density value at a given SF value together with a slope value.
- entering the SF max and min SF values manually to trick the interpolation.
I realize that both these options require negative density values safety checks.
Anyways, back to work. Thank you again for your work on CC!