# Computing Surface Normals¶

Compute normals on a surface.

```# sphinx_gallery_thumbnail_number = 2
from pyvista import examples
import numpy as np
```

Computing the normals of a surface is quite easy using `pyvista.PolyData`’s `pyvista.PolyDataFilters.compute_normals()` method.

```mesh = examples.download_topo_global()
mesh.plot(cmap="gist_earth", show_scalar_bar=False)
``` Now we have a surface dataset of the globe loaded - unfortunately, the dataset shows the globe with a uniform radius which hides topographic relief. Using `pyvista.PolyData.compute_normals()`, we can compute the normal vectors on the globe at all points in the dataset, then use the values given in the dataset to warp the surface in the normals direction to create some exaggerated topographic relief.

```# Compute the normals in-place and use them to warp the globe
mesh.compute_normals(inplace=True)  # this activates the normals as well

# Now use those normals to warp the surface
warp = mesh.warp_by_scalar(factor=0.5e-5)

# And let's see it!
warp.plot(cmap="gist_earth", show_scalar_bar=False)
``` We could also use face/cell normals to extract all the faces of a mesh facing a general direction. In the following snippet, we take a mesh, compute the normals along its cell faces, and extract the faces that face upward.

```mesh = examples.download_nefertiti()
# Compute normals
mesh.compute_normals(cell_normals=True, point_normals=False, inplace=True)

# Get list of cell IDs that meet condition
ids = np.arange(mesh.n_cells)[mesh['Normals'][:,2] > 0.0]

# Extract those cells
top = mesh.extract_cells(ids)

cpos = [(-834.3184529757553, -918.4677714398535, 236.5468795300025),
(11.03829376004883, -13.642289291587957, -35.91218884207208),
(0.19212361465657216, 0.11401076390090074, 0.9747256344254143)]

top.plot(cpos=cpos, color=True)
``` Total running time of the script: ( 0 minutes 15.599 seconds)

Gallery generated by Sphinx-Gallery