M3C2 contour lines and volumetric changes

[daniel_uca]

Hello,
I’m working with the M3C2 algorithm to display erosion on a coastal cliff. I managed to adjust the normals in a way that I get really good results of the changes and am really satisfied.

Now I want to include other measures for the evaluation, apart from the distance.

I’d like to compare the outline of the cliff face before and after the erosion. I computed contour lines of two different point clouds with the “Cross Section” tool by generating “multiple slices”. It is a bit complicated to get the contour lines for the same position on both point clouds but feasible.
I’m having three problems with the contour lines.

1) They are generated as a closed line around the whole point cloud. What I’m interested in is only the cliff face, i.e. the front side of the point cloud, and the double line makes it look weird and confusing.

2) The contour lines of both point clouds are in green colour, which makes a comparison difficult.

3) In areas, where there are holes in the point cloud, the contour line is interpolated as a straight line, which looks bad.

Is there a way to edit the generated contour lines? Can you edit the shape and the colour of the contours? I also have the models as a mesh with filled holes. I think the contour lines would look much better on the mesh but unfortunately the “Cross Section” option is not available for meshes. Is there maybe another way to generate contour lines on a mesh?


Another value that is important as a result of my study, apart from the distance, is the amount of erosion in terms of volume. I read another thread in this forum about this issue of calculating volumetric changes out of M3C2 distances. It seems to be rather inaccurate because of the spaces in between the cylinders of the normals, which are not considered in the calculation.

As I mentioned before, I also have the models in form of meshes and did a comparison with the C2M algorithm. I prefer M3C2 for the distance calculation but maybe there is a way to compute the changes in volume from the comparison of the two meshes. Or maybe you know a different approach for determining the volume of change?

Thanks for your help and for creating this program!

Kind regards,
Daniel

[daniel]

Can you post a snapshot of your cloud and maybe outline on top of it the contour you are expecting to extract? Otherwise, you can still manually extract this (partial) contour with the 'Trace polyline' tool.

And same thing, for the volume you may be able to use the '2.5D Volume Calculation tool' (but it depends on the shape of your clouds).

[daniel_uca]

Hello Daniel,
Thanks for your fast reply!

I tried it again with the "Trace Polyline" Tool. If I use it exactly at the same position for both meshes I get the result I was looking for:



Now the only issue is that I would like to change the colour of the polylines to get a better visibility of the change. Is this possible in Cloud Compare? If not I could probably edit it outside the program.

I tried to estimate the Volume with the "2.5D Volume Calculation" Tool and I get a value. Nevertheless, I'm worried about the accuracy of this value, because only the change in Y direction is considered, which isn't exactly normal to the orientation of the cliff. I think the best way would be to obtain the volume changes from the mesh comparison (C2M algorithm). Do you have an idea how I could solve this problem?

Kind regards!

[daniel]

1) You can select a polyline and then call 'Edit > Colors > Set unique'

2) You can rotate the clouds before using the '2.5D volume calculation tool'.

Otherwise there's no other mean to compute the volume in CC.

[daniel_uca]

Thats perfect! Thanks a lot!

[daniel_uca]

Unfortunately I'm still having problems with the "Trace polyline" tool.
I want to compare polylines from two different models at the exact same position to be able to see the vertical and horizontal changes on the cliff face. After creating the first polyline I toogle the other (eroded) model with the created polyline still visible. I pick the points along the previous polyline to get the same position. What happens is that the algorithm picks the points not only on the model, which is what I want, but also on the first polyline and the result is a chaotic polyline (see screenshot; red = first polyline on the previous model; green = new polyline on the eroded model).



In the manual it says that the tool picks points from a visible cloud or mesh, but this is not the case. I tried it in several ways, but even if the other mesh or other polylines are not visible, points are picked on all entities. The only possiblity to avoid this is to delete all entities from Cloud Compare that should not be considered for the point picking. Nevertheless, doing so I can't reproduce the polyline at the same position.

Do you have an idea for a solution to my problem?
Kind regards!

[daniel]

1) make sure that the vertices of the first polyline are not visible
2) make sure that the picking is CPU-based (and not done with OpenGL). See the 'Auto-compute octree for picking' option in the 'Display options' and set it to 'Always'
3) EDIT: use the 2.8.1 or the latest 2.9.beta version ;)

[daniel_uca]

I installed the stable 2.8.1 version and the "Oversampling" didn't work anymore when creating the Polylines, don't know why. Anyway, I installed the 2.9 Beta and now it's working just the way I wanted. Thank you!