CloudCompare forum

Great! I am going to download the beta right now and check it out.

Thank you,

Lexi

hey i am currently working on my undergraduate thesis using Canupo.
i was wondering if it is possible to produce density plot using this plugins since i am unable to produce the density plot as well using the command prompt.
do any of you have a solution??
thank you very much!

DId you try with the more classical 'Tools > Other > Density' method?

hey daniel!
actually i was hoping to plot the density plot like which to shows the density for each dimensions (i.e 1d 2d 3d)
and is it possible to classify using multiple class ? because my end product only shows two classes.
thanks!

The density plots can't be plotted with CloudCompare (only with the original command line version of Canupo).

And to classify a cloud with more than 2 classes, you need to create several cascading classifiers ( A / not A --> then with the 'not A' points, B / not B, and restart with the 'not B' points, etc.).

im sorry, i do not really get it, if i would classify vegetation, building, and road. what should i do? can you explain the classifier i should create?
thanks

You should cut parts in your point clouds of each class then build one classifier by using the part of the cloud corresponding to your class (road) AND the merge of all others parts (vegetation+building for example) and do this for each class(ifier) you need. Then you can add the whole cloud as a reference when you create your classifier.

Then you open the full cloud you used to create parts and try your classifiers on it, using both SF (class & confidence) to filter your point cloud.

Hope it was usefull :)

hello haowyr!
i am planning to classify road, building, high vegetation and low vegetation.
so should the classifier be :
1. road / building + high vegetation + low vegetation
2. building / high vegetation + low vegetation
3. high vegetation / low vegetation
is this what you mean?
should the road be cut out of the whole data if i want to do number 1 ?
thank you!

I was thinking about (for the building of the classifiers)
- road/building+vegetation (as you cannot split high and low only with classifier)
- building/road+vegetation
- vegetation/building+road

Then you determine the best order to apply (of there is one, by testing) and you apply the first classifier (for example vegetation seems a good start) then you substract the classified part of the cloud from the entire cloud using c2c. You use the result and apply the second classifier, substract etc...

Tell me if it seems clearer to you ;)

available descriptors/scales data mismatch with classifier! compute new descriptor or cancel?

does anyone know what happened? can someone explain to me the process behind the curtain?
pleasee.