Description

This filters implements a robust normal vector estimator for point cloud data.
It can handle sharp features as well as smooth areas.
The method is based on the inclusion of a robust estimator into a Principal Component Analysis in the neighborhood of the studied point which can detect and reject outliers automatically during the estimation.

Parameters

Parameter	Description
Number of neighbors	Number of points taken into account for local surface fitting. Low value leads to a faster computation time but can decrease quality in case of important sampling anisotropy or noise level. Typical: 100. For scattered point clouds: 50. For very noisy point clouds: 300.
Noise (m)	The sensor noise, in meters. For bad sensor: 0.03. For good sensor: 0.005. For artificial point cloud: 0.
Max Curvature (m)	Maximum curvature radius to define a sharp feature, in meters. For curved objects: 0.1. For piecewise planar objects: inf (3.40282e+38).

Supported data structures

The supported data structures are: CGAL Point Set, PCL Point Cloud.

Restrictions

None.

References

Robust normal vector estimation in 3D point clouds through iterative principal component analysis,
Julia Sanchez, Florence Denis, David Coeurjolly, Florent Dupont, Laurent Trassoudaine and Paul Checchin,
ISPRS Journal of Photogrammetry and Remote Sensing, vol. 163, pp. 18-35, 2020.