非常简单的信号处理

Question

I am processing a very simple signal like this:

import numpy as np

series = np.random.random(20)
blind = np.zeros(series.size)

for ii in np.arange(blind.size-1):
    if series[ii+1]-series[ii]>0.4:
        blind[ii] = 1

However, this does not look Numpythonic at all! How can I rewrite the code to use the Numpy-way?

Answer 1

You mean something like ? np.diff(series)>0.4

Answer 2

blind[series[1:] - series[:-1] > 0.4] = 1

or following suggestion from tk

blind[np.diff(series) > 0.4] = 1

The key point being that you use boolean/logical indexing (ie indexing with a boolean array) instead of a loop and an if statement.