subsample - space - min. space between points

[lf2e10]

Hi,
In the "space" method for sub-sampling, a subset of points can be obtained by assigning a minimum distance between points. CloudCompare always produces the same subset (from the same original point cloud) if the same minimum distance is used (run it multiple times and ever time the same subset is produced). Theoretically, if the minimum distance is the only criterion for sub-sampling from the original point cloud, there should be a variety of choices for subset. I guess that there may be some additional constraints so that a unique sub-sample is produced. Does anyone know the algorithm used in CloudCompare for the subsampling by a minimum distance?
Many thanks.
Lei

[daniel]

As the points are always picked in the same order (there's no 'random' process here), then the result will indeed always be the same for a given point cloud and a given 'minimum distance'.

If I remember well, the idea is to take the first point, flag it, then proceed to the next point: if this point is farther from any other flagged point than the minimal distance, then flag it as well (otherwise ignore it). And proceed to the next point, etc. At then end we only keep the flagged points.