UniformGrid.image_dilate_erode(dilate_value=1, erode_value=0, kernel_size=(3, 3, 3), scalars=None, progress_bar=False)#

Dilates one value and erodes another.

image_dilate_erode will dilate one value and erode another. It uses an elliptical footprint, and only erodes/dilates on the boundary of the two values. The filter is restricted to the X, Y, and Z axes for now. It can degenerate to a 2 or 1-dimensional filter by setting the kernel size to 1 for a specific axis.

dilate_valueint or float, optional

Dilate value in the dataset. Default: 1.

erode_valueint or float, optional

Erode value in the dataset. Default: 0.

kernel_sizelist(int) or tuple(int), optional

Length 3 iterable of ints: (xsize, ysize, zsize). Determines the size (and center) of the kernel. Default: (3, 3, 3).

scalarsstr, optional

Name of scalars to process. Defaults to currently active scalars.

progress_barbool, optional

Display a progress bar to indicate progress. Default False.


Dataset that has been dilated/eroded on the boundary of the specified scalars.


This filter only supports point data. Consider converting any cell data to point data using the DataSet.cell_data_to_point_data() filter to convert ny cell data to point data.


Demonstrate image dilate/erode on an example dataset. First, plot the example dataset with the active scalars.

>>> from pyvista import examples
>>> uni = examples.load_uniform()
>>> uni.plot()

Now, plot the image threshold with threshold=[400, 600]. Note how values within the threshold are 1 and outside are 0.

>>> ithresh = uni.image_threshold([400, 600])
>>> ithresh.plot()

Note how there is a hole in the thresholded image. Apply a dilation/ erosion filter with a large kernel to fill that hole in.

>>> idilate = ithresh.image_dilate_erode(kernel_size=[5, 5, 5])
>>> idilate.plot()