Reducing number of triangles

Is there a good way to reduce the number of triangles on a surface? I have a few ways but am curious if there are ways to do this that will keep as much surface data as possible. It would be handy if there was some type of voxel re-mesh.

1 Like

Can you be more specific with what you are trying to accomplish? Are you working with point cloud surfaces? Are your breakline approximation and sampling distance settings negotiable? Are you concerned with maintaining a certain volume?

I am looking to reduce a surface that is created from a point cloud from 4 million triangles down to 1 million triangles. I was curious if there is a way to reduce the number of triangles on the existing surface rather than reducing the point cloud and then creating a surface from the point cloud to achieve it.

The Flatten Surface command would help reduce triangles if you can tolerate its accuracy. I think you’ll like the results better if you manipulate the point cloud though. Either way, you are slightly skewing the surface volume, and using something like the RPS Point Cloud Processor may be more mathematically correct and more customizable for quality control compared to the Flatten Surface command.

Another option could be just reducing the total quantity of points measured in your point cloud surface . In TBC Settings → General → Point clouds you can change max number of points in surface definition
Less points = Less triangles I would think, though I have never worked with point clouds outside of training

1 Like

To keep the vertical integrity of the surface while still decreasing the size would be a 3 step process. First would be to use the Surface to Pts command under the modeling tab, this will create a point cloud from the existing surface. After that, use the point cloud processor command under Tool Shed to decrease the size of the point cloud while keeping the vertical data intact. After that, create a surface from the reduced point cloud.

Using this process should decrease the overall number of vertices in the surface while keeping the vertical data needed.


You can also Grid a surface which will reduce the data size but not intelligently - I would go with Chris Bayliff approach myself if you have our command library - i.e. convert the surface into a Point Cloud and then hit the point cloud with Point Cloud Processor and then model from the resulting point cloud to create your reduced surface. I do this all the time on DOT Highways Projects where the DOT OG is made up of a TIN model derived from a Point Cloud - Often they are way larger than they need to be so just reducing the area of the surface reduces the surface size significantly, if the surface is point cloud based then do the above to strip down the Point Cloud intelligently and then Model from there.