Compute 2.5D volume - understanding

[Lou]

Hello all,

I'm new to CloudCompare and I would like to understand the Compute 2.5D volume tool using as source "constant" and a point cloud. I have a point cloud of a slope that I would like to know the volume. When I compute the volume of this point cloud with the "constant" I have an "added volume" but I don't understand to what is this constant based to calculate this volume and to what this "added volume" corresponds. Can someone help me understand what is this volume ?

image857.png (127.22 KiB) Viewed 6276 times

Thank you very much in advance!

[daniel]

The 2.5D Compute the 'vertical' distances between the points and either another cloud, or a constant plane (Z = constant = 0 in your case). Then it sums the individual volume contribution of each point. As you have set the constant plane as 'after / ceil', the distances are inverted and are probably negative.

And it's 'added volume' if the volume is positive (i.e. the points are above the constant plane) or 'removed' if it's negative.

See https://www.cloudcompare.org/doc/wiki/i ... .5D_Volume

[DA523]

In some cases, 2.5d volume does not work
here is an example

3z.zip

(107.57 KiB) Downloaded 675 times

[daniel]

Yes, hence the '2.5D'. It does not work with real 3D shapes. You cannot have 2 different points with different altitudes at the same XY location.

[antonella.marsico]

Hello all
Is it possible to estimate the accuracy on the rasterization process when using Compute 2.5D volume tool?

Thanks you in advance

[daniel]

Hum, hard to tell. But basically, the algorithm considers that the volume is the distance between 2 horizontal squares that have the size of a grid cell (= diff_height * (cs x cs), where cs is the cell size, and diff_height the difference of height in the 2 equivalent cells of the 'before' and 'after' grids).

So it depends a lot on how you have selected the average height in the cell (min, max, average). But if you had chosen 'min' for the 'before' grid, and 'max' for the 'after' grid, then the error would be at maximum (max - min) * cs^2. Really the worst case of course ;). At least, the smaller cs is, the better (as long as you have enough data to fill the cells!).

[antonella.marsico]

Hi Daniel
thank you very much for your answer.

[antonella.marsico]

Hi Daniel
I tried to follow your suggestion, but I find 'min' for the 'before' grid, and 'max' for the 'after' grid in the "empty cell" setting. I have to compute first the "vertical' distances between each points cloud (by setting the 'min' cell height for before grid and the 'max' cell height for after grid) and a constant plane and then compare the "before" and "after" grid. It is right?

[DA523]

If you have dense pointscloud, apply 2.5D volume
..
but if you do not have dense pointscloud , the steps are :
1- Meshing (to gap between 2 points (linear averaging )
2- If necessary, clean the mesh to be a concave hull not a convex hull (something only humans can do not AI ??)
3- Sample points the mesh (5 million, bigger is better )
4- Apply 2.5 D volume (use leave it empty)

Check attached example

..

Group.zip

(1.03 MiB) Downloaded 361 times

..

sampled.png (253.5 KiB) Viewed 5934 times

[DA523]

In the above example, the volume of pointscloud against constant plane z=0
grid step= 0.055
mesh sample points 10M
volume is 24,279.576 cubic units