machinevisiontoolbox.Image.distance_transform

Image.distance_transform(invert=False, norm='L2', h=1)

Distance transform

Parameters:

  • invert (bool, optional) – consider inverted image, defaults to False

  • norm (str, optional) – distance metric: ‘L1’ or ‘L2’ [default]

  • h (int, optional) – half width of window, defaults to 1

Returns:

distance transform of image

Return type:

Image

Compute the distance transform. For each zero input pixel, compute its distance to the nearest non-zero input pixel.

Example:

>>> from machinevisiontoolbox import Image
>>> import numpy as np
>>> pixels = np.zeros((5,5))
>>> pixels[2, 1:3] = 1
>>> img = Image(pixels)
>>> img.distance_transform().print(precision=3)
 2.324 1.910 1.910 2.324 2.739
 1.369 0.955 0.955 1.369 2.324
 0.955 0.000 0.000 0.955 1.910
 1.369 0.955 0.955 1.369 2.324
 2.324 1.910 1.910 2.324 2.739
>>> img.distance_transform(norm="L1").print()
 3.00 2.00 2.00 3.00 4.00
 2.00 1.00 1.00 2.00 3.00
 1.00 0.00 0.00 1.00 2.00
 2.00 1.00 1.00 2.00 3.00
 3.00 2.00 2.00 3.00 4.00
Note:

  • The output image is the same size as the input image.

  • Distance is computed using a sliding window and is an approximation of true distance.

  • For non-zero input pixels the corresponding output pixels are set to zero.

  • The signed-distance function is image.distance_transform() - image.distance_transform(invert=True)

References:

  • Robotics, Vision & Control for Python, Section 11.6.4, P. Corke, Springer 2023.

Seealso:

opencv.distanceTransform