Distinguish rockwall from vegetation

[boocaj]

Hello,

I have two Lidar dataset of the same Rockwall (one of 2018 and another of 2022). I need to filter the point clouds in order to keep only the rockwalls (remove vegetation) to compare them and analyse if there has been rockfalls in the las years.

I however only have access to a pre-processed dataset for 2018 and the only interesting scalar fields to do so are "intensity" and "classification". "Intensity" does not seem to allow me to distinguish rock and vegetation. "Classification" has 2 different classes which seems to be vegetation (class 1) and ground (class 2). Once I remove the vegetation, I sill have much more than the rockwalls (all ground points).

My question is thus:

Considering the Intensity and the Classifications of the 2018 dataset does not allow me to keep only the rockwalls, is there any way to keep every points that has a certain class value (ground in this case) and that are at more than a certain distance (lets say more than 1m) of another point from another class (vegetation)? Considering there is no vegetation on the rockwalls, only the points corresponding to rock would remain.

FYI: I don't have this problem with the 2022 point cloud because the points intensity allow me filter the points that I need.

Thanks in advance!

[daniel]

So you can technically apply your 'algorithm' by splitting the cloud into 2 classes (with 'Edit > Scalar fields > Filter by Value').

Then you can compute the distances between one and the other (using the 'ground' cloud as the 'compared' cloud).

And eventually you can filter the ground cloud based on the distances value.

(An alternative which might be a little more challenging is to use the Canupo classification plugin: https://www.cloudcompare.org/doc/wiki/i ... O_(plugin))