提取某些行/列值的子矩阵

问题描述

我需要根据行/列索引和切片距离对2D输入数组进行切片。在下面的示例中，我可以从输入矩阵中提取一个3x3的子矩阵，但如果不手动写下索引，我无法将此代码调整为适用于任何我想要的搜索距离：

示例：

import numpy as np

# create matrix
mat_A = np.arange(100).reshape((10, 10))

row = 5
col = 5

# Build 3x3 matrix around the centre point
matrix_three = ((row - 1, col - 1),
                (row, col - 1),
                (row + 1, col - 1),
                (row - 1, col),
                (row, col),  # centre point
                (row + 1, col),
                (row - 1, col + 1),
                (row, col + 1),
                (row + 1, col + 1))

list_matrix_max_values = []

for loc in matrix_three:
    val = mat_A[loc[0]][loc[1]]
    list_matrix_max_values.append(val)


submatrix = np.matrix(list_matrix_max_values)
print(submatrix)

退货：

[[44 54 64 45 55 65 46 56 66]]

例如，如果我想要在行/列索引定义的单元格周围提取一个5x5矩阵，我如何做同样的事情？ 提前感谢！

---

解决方案

S=3 # window "radius"; S=3 gives a 5x5 submatrix
mat_A[row-S+1:row+S,col-S+1:col+S]
#array([[33, 34, 35, 36, 37],
#       [43, 44, 45, 46, 47],
#       [53, 54, 55, 56, 57],
#       [63, 64, 65, 66, 67],
#       [73, 74, 75, 76, 77]])