CloudCompare forum

Hello,

I am looking for a way of extracting the point with the lowest elevation within a given cell size.

For example, I want to use this function to create a bare-earth model of a photogrammetric point cloud captured with a drone. I have read several similar instances in this forum but not exactly what I require.

The Rasterize tool computes a point at the exact cell centre and gives the minimum elevation found within the cell size to that point.

Selecting Resample input cloud within Rasterize selects the point closest to the exact cell centre and assigns the minimum z value to that point (I think).

What I am looking to do is search within an assigned cell size (ex. 5 or 10m) and return the point with the lowest elevation at it's current location. Is this possible?

Any help anyone can provide would be much appreciated. Thanks.

Colin

Looking at the code (to be sure), it seems the Rasterize tool is able to keep the lowest point if you choose the right options.

And that would be: 'Resample input cloud' AND any other option than 'Average' for the cell height.

So 'Resample input cloud' AND 'minimum' cell height should achieve exactly what you are trying to do.

You can double check that by comparing the resulting cloud and the original one. The distances should be 0 for all the resampled points.

I'm also having trouble with this (2.5 years later):

The "resample input cloud" option takes the minimum elevation within the cell, and assigns it to the actual point closest to the cell center. However, this is still not a real observed point, we're changing the elevation of a real point. Since this is survey data, I'd like to end up with an x, y, z point that we actually observed.

Is there really no way to take the actual point with the lowest elevation within a cell of a given size? This seems like a simple enough procedure.

Thanks,
Beau

Which version are you using? Because on my side, with 2.13.alpha, when I use 'minimum' for cell height, and 'resample input cloud' checked, and 'Fill with' = 'leave empty', I get a subset of the original points (and not a mix between XY and min(Z) ).