Find peaks/valleys of a noisy signal

John Williams answered . 2025-07-10 06:08:38

To address your challenge of finding a single peak and valley in a noisy signal, you can try using a smoothing technique to reduce noise before detecting peaks and valleys. Here's an approach using Python and the SciPy library:

1. Smooth the Data: Use a moving average or a Savitzky-Golay filter to smooth the noisy signal.

2. Find Peaks and Valleys: Use the find_peaks function from the scipy.signal module with appropriate parameters.

Here's an example of how to do this:

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy.signal import find_peaks, savgol_filter

# Load your data
data = pd.read_csv('path_to_your_data.csv', header=None)
x = data[0]
y = data[1]

# Smooth the data using Savitzky-Golay filter
window_size = 51  # Choose an appropriate window size
poly_order = 3  # Polynomial order
y_smooth = savgol_filter(y, window_size, poly_order)

# Find peaks
peaks, _ = find_peaks(y_smooth, distance=50)  # Adjust 'distance' as needed

# Find valleys (by finding peaks in the inverted signal)
valleys, _ = find_peaks(-y_smooth, distance=50)  # Adjust 'distance' as needed

# Plot the results
plt.figure(figsize=(10, 6))
plt.plot(x, y, label='Original Signal')
plt.plot(x, y_smooth, label='Smoothed Signal')
plt.plot(x[peaks], y_smooth[peaks], 'ro', label='Peaks')
plt.plot(x[valleys], y_smooth[valleys], 'go', label='Valleys')
plt.legend()
plt.xlabel('Time')
plt.ylabel('Signal')
plt.title('Peak and Valley Detection')
plt.show()

You can adjust the window_size, poly_order, and distance parameters to better suit your data. The window_size should be chosen based on the width of features in your signal, and distance should be set to avoid detecting multiple peaks/valleys around the same point.

Not satisfied with the answer ?? ASK NOW