Extract all minima from an array:

Asked May 27 '19 at 12:17

Active Jul 19 '21 at 21:41

Viewed 55 times

Suppose I have an array in numpy which reads:

arr = np.array([[1,2,7,6],[2,3,2,1],[10,8,1,3],[9,7,4,2], [np.nan, np.nan, np.nan, np.nan]])

I would like to extract all entries where surrounding entries are superior. A sort of 'local minima' such that I obtain the following indices:

indices = [(0,0), [(2,2), (2,3)]]

Note that the second item in the list is a list of tuples. I would like for such an algorithm to be able to identify local minimum areas as well as local minimum points. NAN values should simply be ignored and treated like a "border".

I have tried playing around with np.gradient because I thought that it should allow me to extract parts where the derivative is 0. But I couldn't really get it to a point where it would work.

Do any libraries or out-of-the-box functionalities exist to achieve this?

edited Jul 19 '21 at 21:41

halfer

19,471
17
87
173

asked May 27 '19 at 12:17

user32882

4,174
4
28
56

surroundings, as in both axis? – yatu May 27 '19 at 12:18
exactly. Both axes are important. If there are any NAN values I also want them to act as a border. For instance `np.array([[np.nan, np.nan, np.nan], [8,1,3], [7,4,2]]) `should yield`[(1,1)]`. – user32882 May 27 '19 at 12:21
https://stackoverflow.com/questions/4624970/finding-local-maxima-minima-with-numpy-in-a-1d-numpy-array – Aly Hosny May 27 '19 at 14:07
@Aly Hosny That's for a 1D array. That's easy... – user32882 May 28 '19 at 03:12

Extract all minima from an array:

0 Answers0