Page 1 of 1

Comparison of point cloud classifications

Posted: Sat Mar 25, 2017 9:37 am
by tyna77
Hi,
I need to compare two the same point clouds which has different classifications. One of the classification is manual and I take this one as my reference to compare with other. I have three classes there. Buildings, vegetation and ground. I found out that the comparison does not work with the whole clouds, because it’s the same cloud, so I decided to compare two ground classes and two building classes separately. I have some results but I am not sure if it is right or wrong because the point clouds are not completely overlapped.

Thanks for any advice or idea

Kristýna

Re: Comparison of point cloud classifications

Posted: Mon Mar 27, 2017 7:18 pm
by daniel
You can indeed compute the distances between two different subsets of the same cloud (i.e. two equivalent classes obtained with different methods). But in this case the only interesting measure would be the number of points in each class? You don't really need to compute the distances (apart if you are interested to see how 'far' one method overflows?

Re: Comparison of point cloud classifications

Posted: Wed Mar 29, 2017 5:08 pm
by tyna77
Thank you for your rensponse. That’s exactly what I need, to see how far one point cloud overflows the second one.

And I have one more question. In the manual I read that for clouds that have low density or holes should be used local models. My clouds belong to this case. But I‘m not sure which method I should use. Can I use Quadric when my clouds represent ground?

Re: Comparison of point cloud classifications

Posted: Wed Mar 29, 2017 5:20 pm
by daniel
Yep, should be fine.

Re: Comparison of point cloud classifications

Posted: Wed Mar 29, 2017 7:01 pm
by tyna77
Thanks a lot for your help.