本文实例讲述了python3实现的旋转矩阵图像算法。分享给大家供大家参考,具体如下:
问题:
给定一个 n × n 的二维矩阵表示一个图像。
将图像顺时针旋转 90 度。
方案一:先按x轴对称旋转, 再用zip()解压,最后用list重组。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
|
# -*- coding:utf-8 -*-
#! python3
class solution:
def rotate( self , matrix):
"""
:type matrix: list[list[int]]
:rtype: void do not return anything, modify matrix in-place instead.
"""
matrix[:] = map ( list , zip ( * matrix[: : - 1 ]))
return matrix
if __name__ = = '__main__' :
# 测试代码
matrix = [
[ 1 , 2 , 3 , 4 ],
[ 5 , 6 , 7 , 8 ],
[ 9 , 10 , 11 , 12 ],
[ 13 , 14 , 15 , 16 ]
]
solution = solution()
result = solution.rotate(matrix)
print (result)
|
运行结果:
[[13, 9, 5, 1], [14, 10, 6, 2], [15, 11, 7, 3], [16, 12, 8, 4]]
方案二:找到规律,用原矩阵数据 赋值
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
|
# -*- coding:utf-8 -*-
#! python3
class solution:
def rotate( self , matrix):
"""
:type matrix: list[list[int]]
:rtype: void do not return anything, modify matrix in-place instead.
"""
m = matrix.copy()
n = len (matrix)
for i in range (n):
matrix[i] = [m[j][i] for j in range (n - 1 , - 1 , - 1 )]
return
if __name__ = = '__main__' :
# 测试代码
matrix = [
[ 1 , 2 , 3 , 4 ],
[ 5 , 6 , 7 , 8 ],
[ 9 , 10 , 11 , 12 ],
[ 13 , 14 , 15 , 16 ]
]
solution = solution()
result = solution.rotate(matrix)
print (result)
|
运行结果:
[[13, 9, 5, 1], [14, 10, 6, 2], [15, 11, 7, 3], [16, 12, 8, 4]]
希望本文所述对大家python程序设计有所帮助。
原文链接:https://blog.csdn.net/zhenghaitian/article/details/80935434