Poisson Surface Reconstruction Reference

Pierre Alliez, Laurent Saboret, Gaël Guennebaud

This package implements a surface reconstruction method: Poisson Surface Reconstruction. It takes as input a set of points with oriented normals and computes an implicit function. The CGAL surface mesh generator can then be used to extract an iso-surface from this function.

Introduced in: CGAL 3.5
Depends on: CGAL and Solvers
BibTeX: cgal:asg-srps-18b
License: GPL
Windows Demo: Polyhedron demo
Common Demo Dlls: dlls

Classified Reference Pages

Classes

• CGAL::Poisson_reconstruction_function<GeomTraits>

Classes
class	CGAL::Poisson_reconstruction_function< Gt >
	Implementation of the Poisson Surface Reconstruction method. More...

Functions
template<typename PointInputIterator , typename PointMap , typename NormalMap , typename PolygonMesh , typename Tag = CGAL::Manifold_with_boundary_tag>
bool	CGAL::poisson_surface_reconstruction_delaunay (PointInputIterator begin, PointInputIterator end, PointMap point_map, NormalMap normal_map, PolygonMesh &output_mesh, double spacing, double sm_angle=20.0, double sm_radius=30.0, double sm_distance=0.375, Tag tag=Tag())
	Performs surface reconstruction as follows: More...

Function Documentation

template<typename PointInputIterator , typename PointMap , typename NormalMap , typename PolygonMesh , typename Tag = CGAL::Manifold_with_boundary_tag>

bool CGAL::poisson_surface_reconstruction_delaunay	(	PointInputIterator	begin,
		PointInputIterator	end,
		PointMap	point_map,
		NormalMap	normal_map,
		PolygonMesh &	output_mesh,
		double	spacing,
		double	sm_angle = 20.0,
		double	sm_radius = 30.0,
		double	sm_distance = 0.375,
		Tag	tag = Tag()
	)

Performs surface reconstruction as follows:

• compute the Poisson implicit function, through a conjugate gradient solver, represented as a piecewise linear function stored on a 3D Delaunay mesh generated via Delaunay refinement
• meshes the function with a user-defined precision using another round of Delaunay refinement: it contours the isosurface corresponding to the isovalue of the median of the function values at the input points
• outputs the result in a polygon mesh

This function relies mainly on the size parameter spacing. A reasonable solution is to use the average spacing of the input point set (using compute_average_spacing() for example). Higher values increase the precision of the output mesh at the cost of higher computation time.

Parameters sm_angle, sm_radius and sm_distance work similarly to the parameters of SurfaceMeshFacetsCriteria_3. The latest two are defined with respect to spacing.

Template Parameters

PointInputIterator	is a model of InputIterator.
PointMap	is a model of ReadablePropertyMap with value type Point_3<Kernel>.
NormalMap	is a model of ReadablePropertyMap with value type Vector_3<Kernel>.
PolygonMesh	a model of MutableFaceGraph with an internal point property map.
Tag	is a tag whose type affects the behavior of the meshing algorithm (see make_surface_mesh()).

Parameters

begin	iterator on the first point of the sequence.
end	past the end iterator of the point sequence.
point_map	property map: value_type of InputIterator -> Point_3.
normal_map	property map: value_type of InputIterator -> Vector_3.
output_mesh	where the reconstruction is stored.
spacing	size parameter.
sm_angle	bound for the minimum facet angle in degrees.
sm_radius	bound for the radius of the surface Delaunay balls (relatively to the average_spacing).
sm_distance	bound for the center-center distances (relatively to the average_spacing).
tag	surface mesher tag.

Returns

true if reconstruction succeeded, false otherwise.

#include <CGAL/poisson_surface_reconstruction.h>

Examples:

Poisson_surface_reconstruction_3/poisson_reconstruction_function.cpp.