warp.geometry#
Geometry processing operations.
This module provides functions to process 2D and 3D geometry and their associated topological data structures (e.g., meshes), and GPU-accelerated isosurface extraction.
IsoSurfaceBase defines the interface for dense-grid isosurface
extraction. IsoSurfaceMarchingCubes implements it and extracts a
triangular mesh from a 3-D warp.float32 field sampled at grid nodes.
Sparse extraction skips the dense grid entirely. sparse_marching_cubes()
takes an implicit function and builds a Lipschitz octree around the level set,
so cost scales with surface area rather than volume. lipschitz_octree()
and sparse_marching_cubes_from_cells() expose its two stages separately.
The Python API includes array-level functions, such as swept_volume_mesh(),
that launch kernels over a whole mesh or grid. The kernel API includes Warp
functions, such as swept_volume_sdf(), that evaluate a single point and may
be called from your own warp.kernel() definitions and other Warp functions.
- Usage:
This module must be explicitly imported:
import warp.geometry
Python API#
Abstract base class for isosurface extraction from dense 3D scalar fields. |
|
A reusable context for marching cubes surface extraction. |
|
Method used to classify inside/outside when evaluating the per-mesh SDF. |
|
Flip non-Delaunay interior edges in place until all are Delaunay or max passes is reached. |
|
Return max-depth octree cells that may intersect a level set of a Lipschitz-bounded implicit field. |
|
Extract an isosurface from an implicit function using a Lipschitz octree. |
|
Extract an isosurface from an explicit set of occupied cells and corner values. |
|
Return the world-space axis-aligned bounds of every mesh over every pose. |
|
Compute a dense regular-grid discretization of the swept-volume pseudo signed-distance function. |
|
Extract the swept volume (motion envelope) of animated rigid meshes. |
|
Build triangle-triangle adjacency for a triangle mesh. |
Kernel API#
Return the local edge of |
|
Evaluate the swept-volume pseudo signed distance at a single point. |