scalar fields getting damaged

[PablerasBCN]

I've not yet detected the moment this happens, but the scalar fields are getting damaged.

I believe it occurs when merging clouds. As scalar fields get values lost in scattered parts of clouds. and I'm segmenting clouds and merging all the time atm.

I also noticed and old bug thaht is thath scalar fields are listed several times when opening the file. I'll report further as I go., thankfully I can transfer the data lost from backups.

[daniel]

Is it with the latest 2.14.alpha version?

[daniel]

If you find a way to reproduce it on your side, I'm interested (because right now I haven't been able to reproduce it on my side)

[PablerasBCN]

In the latest version of yes, the one that was pushed so the corrupt.bin was possible to read.

I found a bug, that could be related to the issue.

When you merge 2 clouds and you tell to YES generate an scalar field so you can ID each. It is not working, it only gets one color.

Not sure if it matters, but the data is in a .bin. I could send the file if helps. May be do a crop so it is not too heavy. let me know

[daniel]

Ok, I believe I've found the bug (it was impacting this 'merge index' feature, the Closest Point Set calculation, the 'check points inside frustum' feature, point visibility computation, and the 3Dmasc classification field... Hopefully nothing else.

I've released a new 2.14.alpha version online.

[PablerasBCN]

Hello Daniel,

I tested with alpha 2024/08/12 version but the issue persists!

I sen´t you email to
admin@cloudcompare.org

With small sample dataset, when you merge the two clouds, you'll see that no index for each cloud is generated.

Cheers!

[daniel]

Yes, I uploaded the wrong version! This is now fixed.

[PablerasBCN]

I think I noticed what could be going on on the damaged SC.

This is in the latest version, the point cloud Indx ID, may have been a different but.

I was already having this impression, but I saw it happen right in front of my eyes this time.

I'm accuratelly segmenting ground/ off ground.

So, there was an small chunk of asphalt in the non ground, and UPON MERGING, with the ground cloud I could VERY CLEARLY, see still the chunk despite I was visualizing INTENSITY sc, which would need it to clearly blend.

I did even now open the original tile, and indeed there is nothing in the intensity.

It feels like the SC values are merged with a shift, based on the active dynamic scale. I had "grey" in both, before merging.

[PablerasBCN]

also notice how the intensity values in the graph go down to -5k now


ORIGINAL


AFTER MERGE

[PablerasBCN]

I tried to reproduce using the same cloud but It is not as simple as segment out, and merging.... sadly.....