Hi,
issue:
the mode parameter seems to be referenced both in the left input measure and in the switch node. When you select ‘area’ mode, it would trigger the left input measure, that is set to perimeter. I don't understand the logic.

proposal:
The current threshold method can be helpful when you know you might end up with different disconnected pieces that you want to preserve some of them.
I believe the most common workflow is that you work with one single piece and you just want to get rid of all the small disconnected garbage created by the volume. So adjusting the threshold can be time consuming and not really procedural.

I currently use this setup instead: I promote the class attribute with the method “mode”, in order to know the most common match (which is of course the class of the biggest piece with most primitives). Then I simply blast the primitives whose class don't match the promoted class.
Maybe it's worth to add this option?

Cheers

Hi Andr,

That switch does seem odd. I will look into this.
I also appreciate your suggestion, which makes a lot of sense.

What about a toggle called “Keep largest piece”? Largest number of primitives to determine biggest piece is not the most robust solution for that though. I'd rather sum op the values per connected piece and use that data. (Or allow user to pick between sum of measured area/perimeter and number of primitives)

Paul

Hi,

I have fixed the reported issue, and have implemented your RFE.
Let me know what you think!

Paul

Hey Paul thanks a lot for the effort!
will check it out as soon .172 becomes avail.

cheers