poisson

Module Contents

Functions

mesh_from_oriented_points(points, normals[, ...])

Creates a triangle mesh from an oriented point cloud using Poisson surface reconstruction.
poisson.mesh_from_oriented_points(points, normals, octree_depth=0, samples_per_node=1.5, interpolation_weight=2.0, use_normal_length_as_confidence=False, use_dirichlet_boundary=False, colors=None, output_vertex_depth_attribute_name='', verbose=False)

Creates a triangle mesh from an oriented point cloud using Poisson surface reconstruction.

Parameters:

  • points (Annotated[numpy.typing.ArrayLike, dict(dtype='float64', order='C', device='cpu')]) – Input point cloud positions (N x 3 matrix).

  • normals (Annotated[numpy.typing.ArrayLike, dict(dtype='float64', order='C', device='cpu')]) – Input point cloud normals (N x 3 matrix).

  • samples_per_node (float) – Number of samples per node.

  • octree_depth (int) – Maximum octree depth. (If the value is zero then log base 4 of the point count is used.)

  • interpolation_weight (float) – Point interpolation weight (lambda).

  • use_normal_length_as_confidence (bool) – Use normal length as confidence.

  • use_dirichlet_boundary (bool) – Use Dirichlet boundary conditions.

  • colors (Annotated[numpy.typing.ArrayLike, dict(order='C', device='cpu')] | None) – Optional color attribute to interpolate (N x K matrix).

  • output_vertex_depth_attribute_name (str) – Output density attribute name. We use a point’s target octree depth as a measure of the sampling density. A lower number means a low sampling density, and can be used to prune low-confidence regions as a post-process.

  • verbose (bool) – Output logging information (directly printed to standard output).

Return type:

lagrange.core.SurfaceMesh