福清建设局网站深圳seo技术

看代码看见这个方法，记录一下，这个是人家官网的链接np.correlate

云里雾里的，其实就是两个数组点乘，不同模式就是错位点乘，直接看代码

a是原本的数组，v就是滤波器，对应相乘

import numpy as npmode0 = 'same'
mode1 = 'valid'
mode2 = 'full'a = [1, 2, 3]
v = [1, 2]print(np.correlate(a, v, mode0))
print(np.correlate(a, v, mode1))
print(np.correlate(a, v, mode2))# 结果
[2 5 8]
[5 8]
[2 5 8 3]

a = [1, 2, 3]
v = [1, 2, 3]print(np.correlate(a, v, mode0))
print(np.correlate(a, v, mode1))
print(np.correlate(a, v, mode2))# 结果
[ 8 14  8]
[14]
[ 3  8 14  8  3]

a = [1, 2, 3, 4]
v = [1, 2]print(np.correlate(a, v, mode0))
print(np.correlate(a, v, mode1))
print(np.correlate(a, v, mode2))# 结果
[ 2  5  8 11]
[ 5  8 11]
[ 2  5  8 11  4]