There are functions in scipy.signal module that can be used to find local minima and maxima in an array/matrix.
Function argrelmin() is used to calculate the relative minima of data. Similarly, argrelmax() is used to calculate the relative maxima of data. These functions return the indices of local minima/maxima in the data. The indices can be used to find the values of local minima/maxima.
The following example shows how to use these functions. I am also plotting the data so that the local minima/maxima can be verified visually.
import numpy as np
from scipy.signal import argrelmin, argrelmax
import matplotlib.pyplot as plt
# generate random data
np.random.seed(7)
y = np.random.random_sample(25)
x = np.asarray([i+1 for i in range(25)])
# find local minima and maxima
min_idx = argrelmin(y)
print("Indices of minima: {0}".format(min_idx))
print("Minima: {0}".format(y[min_idx]))
max_idx = argrelmax(y)
print("Indices of maxima: {0}".format(max_idx))
print("Maxima: {0}".format(y[max_idx]))
# plot data
plt.plot(x, y, 'g-o')
plt.savefig("plots/local_minima_maxima.png")
The above code prints the following output:
Indices of minima: (array([ 2, 7, 13, 16, 19, 22]),)
Minima: [0.43840923 0.07205113 0.06593635 0.21338535 0.02489923 0.23030288]
Indices of maxima: (array([ 1, 4, 11, 15, 18, 21]),)
Maxima: [0.77991879 0.97798951 0.80373904 0.90959353 0.93120602 0.9501295 ]
It generates the following plot. You can verify the results by looking at the plot.
