0

Let's say I have a 2D array with a size of m x n elements. Now, I want to get the indices of all maximums. So the result should be something like: [(m1, n1), (m2, n2)] where m and n indicate the x and y coordinates of my maximums.

With only one maximum its quite easy, but with more, I'm getting stuck.

import numpy as np

pixel = np.array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
                  [0, 189, 12, 0, 0, 1, 0, 0, 0, 0],
                  [0, 6, 0, 0, 0, 0, 0, 0, 0, 0],
                  [0, 0, 0, 1, 0, 0, 0, 203, 9, 0],
                  [0, 0, 0, 0, 0, 0, 0, 12, 0, 0],
                  [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
                  [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
                  [0, 0, 5, 245, 0, 0, 0, 7, 4, 0],
                  [0, 0, 0, 0, 0, 0, 0, 250, 8, 0],
                  [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]])
result = np.where(pixel == pixel.max())

print("cross detection at y:", result[0][0], "x:", result[1][0])

print(pixel)

Does somebody have an idea? It would be great, thanks!

normanius
  • 8,629
  • 7
  • 53
  • 83
Bakira
  • 79
  • 9
  • can you please define what are those maximums ?. – pymym213 Jan 19 '20 at 11:02
  • 1
    Does this answer your question? [How to make numpy.argmax return all occurrences of the maximum?](https://stackoverflow.com/questions/17568612/how-to-make-numpy-argmax-return-all-occurrences-of-the-maximum) – Mohammed Deifallah Jan 19 '20 at 11:03
  • 1
    Your question is not clear. – s.ouchene Jan 19 '20 at 12:23
  • 2nd try: basically I have an image which contains some blurred spots and I want to detect the pixel, where these spots are. For example, I have 4 spots, than I expect a list with 4 x,y coordinates. Because of the imaging structure, I know that two points cannot be closer than 20 pixels. Maybe now it gets clearer.. – Bakira Jan 19 '20 at 20:54

2 Answers2

1

Try this:

x,y = np.where(pixel == np.max(pixel))

this will return x axis and y axis of all the elements with maximum values Now,for your question you can do

np.array((x,y)).T

for this last code I look up to this question

Shubham Shaswat
  • 1,250
  • 9
  • 14
0

Try numpy.argmax, it returns the indices of the maximum values along an axis.

Stoufa
  • 23
  • 7