曼哈顿距离的距离变换-Python/NumPy/SciPy [英] Distance transform with Manhattan distance - Python / NumPy / SciPy

问题描述

我想使用Python和Numpy生成一个2d数组:

I would like to generate a 2d Array like this using Python and Numpy:

[
  [0, 1, 2, 3, 4, 4, 3, 4],
  [1, 2, 3, 4, 4, 3, 2, 3],
  [2, 3, 4, 4, 3, 2, 1, 2],
  [3, 4, 4, 3, 2, 1, 0, 1],
  [4, 5, 5, 4, 3, 2, 1, 2]
]

几乎所有的数字从零开始向左和向右传播.该矩阵允许查看任何点到最接近零的距离.我以为这个矩阵很普通，但是我在网上找不到任何东西，甚至它的名字也没有.如果您有代码可以有效地生成这样的矩阵，或者至少知道如何调用它，请告诉我.

Pretty much the the numbers spread left and right starting from the zeros. This matrix allows to see the distance of any point to the closest zero. I thought this matrix was common, but I couldn't found anything on the web, even its name. If you have a code to efficiently generate such a matrix or know at least how it's called, please let me know.

谢谢

推荐答案

这里是带有

Here's one with Scipy cdist -

from scipy.spatial.distance import cdist

def bwdist_manhattan(a, seedval=1):
    seed_mask = a==seedval
    z = np.argwhere(seed_mask)
    nz = np.argwhere(~seed_mask)

    out = np.zeros(a.shape, dtype=int)
    out[tuple(nz.T)] = cdist(z, nz, 'cityblock').min(0).astype(int)
    return out

在MATLAB中，它称为 Distance transform of binary image ，因此是派生的名字在这里给出.

In MATLAB, it's called Distance transform of binary image, hence a derivative name is given here.

样品运行-

In [60]: a # input binary image with 1s at "seed" positions
Out[60]: 
array([[1, 0, 0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0, 0, 1, 0],
       [0, 0, 0, 0, 0, 0, 0, 0]])

In [61]: bwdist_manhattan(a)
Out[61]: 
array([[0, 1, 2, 3, 4, 4, 3, 4],
       [1, 2, 3, 4, 4, 3, 2, 3],
       [2, 3, 4, 4, 3, 2, 1, 2],
       [3, 4, 4, 3, 2, 1, 0, 1],
       [4, 5, 5, 4, 3, 2, 1, 2]])

这篇关于曼哈顿距离的距离变换-Python/NumPy/SciPy的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！