Surface Smoothing

Smoothing rough edges of a surface mesh

# sphinx_gallery_thumbnail_number = 4
from pyvista import examples

Suppose you extract a volumetric subset of a dataset that has roughly defined edges. Perhaps you’d like a smooth representation of that model region. This can be achieved by extracting the bounding surface of the volume and applying a pyvista.PolyData.smooth() filter.

The below code snippet loads a sample roughly edged volumetric dataset:

# Vector to view rough edges
cpos = [-2, 5, 3]

# Load dataset
data = examples.load_uniform()
# Extract a rugged volume
vol = data.threshold_percent(30, invert=1)
vol.plot(show_edges=True, cpos=cpos)

Out:

[(-5.270462178580788, 28.92615544645197, 19.155693267871182),
 (4.5, 4.5, 4.5),
 (0.0, 0.0, 1.0)]

Extract the outer surface of the volume using the pyvista.DataSetFilters.extract_geometry() filter and then apply the smoothing filter:

# Get the out surface as PolyData
surf = vol.extract_geometry()
# Smooth the surface
smooth = surf.smooth()
smooth.plot(show_edges=True, cpos=cpos)

Out:

[(-5.270462173884933, 28.926155439895044, 19.155693264529337),
 (4.500000001480775, 4.500000001480775, 4.500000001480775),
 (0.0, 0.0, 1.0)]

Not smooth enough? Try increasing the number of iterations for the Laplacian smoothing algorithm:

# Smooth the surface even more
smooth = surf.smooth(n_iter=100)
smooth.plot(show_edges=True, cpos=cpos)

Out:

[(-5.2703475762355385, 28.925595495391537, 19.155326046354773),
 (4.499921872800769, 4.499921872800769, 4.499921872800314),
 (0.0, 0.0, 1.0)]

Still not smooth enough? Increase the number of iterations for the Laplacian smoothing algorithm to a crazy high value:

# Smooth the surface EVEN MORE
smooth = surf.smooth(n_iter=1000)
smooth.plot(show_edges=True, cpos=cpos)

Out:

[(-5.015056085827215, 28.092312435289877, 18.633064286399282),
 (4.444192063063383, 4.444192063063383, 4.444192063063383),
 (0.0, 0.0, 1.0)]