Calculated volume in 2.5 d changes with grid step size bug?

[creature]

I tried the latest version cloud compare and i think i might have found a new bug.

So im back at doing a stockpile calculation. I don't get the 2 point clouds error anymore. YaY

Now this happens:

When i change grid step size the calculated volumes are changing by like 3 times to lets say 9 times its calculated value.

Back on version 2.6.3 i don't have this issue.

[daniel]

I tested with the 2.10.alpha version and it works without such an issue.

Are you sure you don't get a warning about 'at least one cloud being sparse'? Typically because the smallest grid step is too small and empty cells appear?

[creature]

I am using a single point cloud to calculate the volume.
I am lost ?
am i missing something in new version?
I used 2.9.1 not 2.9 as pic says.
I get same in numbers 2.10 alpha
Version 2.6.3 is really close when judging by the amount of truck loads used.
I do indeed get sparse warning in new version but works great in older version and is really accurate to real volume.
I didnt save the work but the pile consists out of about 334945 points

old new.jpg (238.86 KiB) Viewed 7252 times

[daniel]

Well, the 'sparse' grid warning explains it all (I mean the wrong volume). Why is it working in the 2.6 version is the question... Can you send me the cloud maybe? (admin [at] cloudcompare.org).

One possibility is that there was a bug related to the grid step in the older version?! Anyway I guess you should still have a good volume with a step twice or thrice bigger than the current one (considering the grid size).

[creature]

Here is the point cloud
The point cloud might vary slightly from the original as that one wasn't saved.
It originates from the same point cloud that i cut it out off.

Pile.ply

google drive link
https://drive.google.com/file/d/15SKnLm ... sp=sharing

[daniel]

Hum, now I wonder if the bug was not in the 2.9 version! Because with the latest 2.10.alpha version, I get almost the same result as with the 2.6 one :D (well I forgot to activate the interpolation mode, but the result has the same magnitude at least).

cc_volume_calc.jpg (159.44 KiB) Viewed 7200 times

I realize now that the result you get with the 2.9 version is very weird because despite the interpolation you get this 'sparse grid' error (which shouldn't happen in this case!).

So you should definitely use the 2.10 version.

[creature]

Humm... i am at a total loss
I tried the 2.10 latest release and i did the same thing with the Pile cloud again.
I cannot produce your results.

I get again the sparse error and a totally wrong result again.
Preview shows dots instead of a nice smooth clouds surface

Does this get calculated on the GPU?
Running on: AMD RX580 GPU, CPU AMD RYZEN 7 1700 "latest drivers installed"
Edit: on March 9th i tried it on a different pc. I'm still getting sparse cloud error.

[daniel]

I finally found the issue! The strategy to fill the empty cells for the 'ceil' cloud was ignored. Only the 'ground' one was used.

I've updated the online 2.10.alpha version.

Thanks for the feedback ;)

[creature]

Thank you for all your amazing hard work!
It's working now!