I'm working on 2D array analysis or an image analysis and get to know the different block of peak values marked "circle" in attached plot [e.g. 2*2 matrix] of the 2D array [as shown in Fig. 01]. Fig. 01: Peaks 2d array

The image formation of the plot look like as Fig. 02 Image plot of Fig. 01

I visited the link https://stackoverflow.com/questions/3684484/peak-detection-in-a-2d-array/3689710#3689710 and use the following script

#NOTE: I get the code from the link above
from scipy import *
from operator import itemgetter  
n = 3  # how many fingers are we looking for
#d = loadtxt("paw.txt")
#data = [DATA of a 2d array]
width, height = data.shape
# Create an array where every element is a sum of 2x2 squares.
fourSums = data[:-1,:-1] + data[1:,:-1] + data[1:,1:] + data[:-1,1:]

# Find positions of the fingers.
# Pair each sum with its position number (from 0 to width*height-1),
pairs = zip(arange(width*height), fourSums.flatten())
# Sort by descending sum value, filter overlapping squares

def drop_overlapping(pairs):
    no_overlaps = []
    def does_not_overlap(p1, p2):
        i1, i2 = p1[0], p2[0]
        r1, col1 = i1 / (width-1), i1 % (width-1)
        r2, col2 = i2 / (width-1), i2 % (width-1)
        return (max(abs(r1-r2),abs(col1-col2)) >= 2)
    for p in pairs:
        if all(map(lambda prev: does_not_overlap(p,prev), no_overlaps)):
    return no_overlaps

pairs2 = drop_overlapping(sorted(pairs, key=itemgetter(1), reverse=True))

# Take the first n with the heighest values
positions = pairs2[:n]

# Print results print(data, "\n")
for i, val in positions:
    row = i // (width-1)
    column = i % (width-1)
    print("sum = %f @ %d,%d (%d)" % (val, row, column, i))

The above script provides 2 adjacent matrix[2*2] of the same peaks.

## 01: Local matrix of the peak
sum = 0.002482 @ 128,190 (65342)
[[0.0006202  0.00062187]
 [0.00061926 0.00062093]]
## 02: Local matrix of the peak
sum = 0.002479 @ 128,191 (65343)
[[0.00062187 0.0006184 ]
 [0.00062093 0.00061746]]

Also, I've tried with varying 'n' and no. of rows and columns but it gives me different local matrix for the same peak (probably I missed something) which does not match with my expectation. I want to find local peaks or a number of local maxima as shown in fig. 01 and Fig. 02. How can I get the local maxima efficiently? Can anybody help me in this regard?

  • $\begingroup$ Do you mean finding a pair of indices, $n, m$ so that $y[n,m] > y[n\pm1, m\pm1]$? $\endgroup$ Commented Feb 19, 2022 at 8:47
  • $\begingroup$ Yes, a pair of indices with data (inside the 2 blobs but I got pair of indices inside 1 blob). In Fig. 02, I want to find two local maxima of size [P*Q] from the 2 red blobs encircled. However, I got the 2 maxima of size [2*2] at location (128,190) & (128, 191) that represents 2 maximum data block but located in the same blob. $\endgroup$
    – CEB
    Commented Feb 19, 2022 at 10:12
  • $\begingroup$ Please don’t cross-post. stackoverflow.com/q/71175329/7328782 $\endgroup$ Commented Feb 19, 2022 at 15:53

1 Answer 1


Look at the 1d function findpeaks. It is well documented.
All needed is to implement those simple ideas in 2d.
Mostly the minimum distance between 2 peaks.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.