Maximum circle inside a polygon

Replies to This Discussion

Permalink Reply by Daniel Piker on September 12, 2017 at 8:20am

This old discussion might help: http://www.grasshopper3d.com/forum/topics/maximum-inscribed-circle

I think solving this exactly for general cases is not trivial, but practically it should be fairly easy to find a good enough solution.

Permalink Reply by Jean-Louis Herman on September 12, 2017 at 9:16am

Well i read this discussion and i did not found the solution. It seems that the circle computed are not inside but around the polygon, as in gen_max_circle_2_polyline.gh definition.

Attachments:

• max circle around a polygon.PNG, 154 KB
• gen_max_circle_2_polyline.gh, 13 KB

Permalink Reply by Daniel Piker on September 12, 2017 at 11:26am

The approach Mateusz showed in that discussion seems to work fine.

Accuracy is determined by how finely you sample your curve for the Delaunay.

Attachments:

• max_inscribed.gh, 9 KB

Permalink Reply by Laurent DELRIEU on September 12, 2017 at 1:25pm

For a polygon you could use crest of medial axis, the heighest crest could be place where there is the biggest circle. 

http://www.grasshopper3d.com/forum/topics/sand-dune-medial-axis-sam...

For general curves the approximation of medial axis using Voronoi could be another to do it. There are many discussions on that.

Permalink Reply by Laurent DELRIEU on September 12, 2017 at 1:32pm

Here I filter Vertex with their Z value the highest is used to get the the point where the maximum circle inscribed is. The radius of circle is depending on the height and on the slope of the roof. 

Et voilà

PS : My script has some problem with 1 curve and work more well with 2 or more. A tree problem  :-(

Attachments:

• Sand_dune_medial_axis_closed_multiple_polygons_v4.gh, 43 KB

Permalink Reply by Laurent DELRIEU on September 12, 2017 at 1:58pm

Choose constant slope. As there are some bad results in brep calcultions look at the roof brep to verify that everything is OK.