maximum point cloud comparison

[lidargeek]

Daniel,

Great software! Thanks for sharing with everyone.

I'm looking at comparing two fairly large lidar scans of a landscape feature. The feature is moving at times a meter per hour. The point cloud is very dense and the features are easy to see from scan to scan. My question, is there a point in cloudCompare where the distance is too great to compare two point clouds? say 1m or 50m?

If there is a limit, maybe there is a better approach to compare data with this much change?

[daniel]

Hi, there's no limitation in the distance computation algorithm (and it's quite a problem sometimes: if both entities are very far - e.g. typically more than 50% of their bounding box max dimension - the processing time can increasing a lot (the farther is the nearest point, the more cells of the octree must be tested and generally the more neighbors must be tested also).

The user can set its own limit (with the 'max dist.' field) in which case all points above this distance will be tagged with this max dist. value then skipped).

[lidargeek]

Thanks Daniel for the prompt reply.

Can you speak to calculating displacement in multiple dimensions? IN particular, we are looking at a feature that is flowing horizontally and shows vertical change from cloud to cloud. Any recommendations on how to best characterize this using cloudCompare?

[daniel]

CloudCompare only computes distances to the nearest point/triangle in 3D. There's no way (for now) to add any particular constraint on the direction of the 'closest' element.

However, one can split the resulting scalar field into 3 sub-fields (one for each dimension X,Y and Z) with the 'split X,Y and Z components' checkbox (for cloud/cloud distances only). They simply correspond to the displacement vectors coordinates. So maybe you can only keep the 'Z' one in your case?

Best