Develop Reference

How to find period of signal (autocorrelation vs fast fourier transform vs power spectral density)?

Suppose one wanted to find the period of a given sinusoidal wave signal. From what I have read online, it appears that the two main approaches employ either fourier analysis or autocorrelation. I am trying to automate the process using python and my usage case is to apply this concept to similar signals that come from the time-series of positions (or speeds or accelerations) of simulated bodies orbiting a star.
For simple-examples-sake, consider x = sin(t) for 0 ≤ t ≤ 10 pi.
import numpy as np
from scipy import signal
import matplotlib.pyplot as plt
## sample data
t = np.linspace(0, 10 * np.pi, 100)
x = np.sin(t)
fig, ax = plt.subplots()
ax.plot(t, x, color='b', marker='o')
ax.grid(color='k', alpha=0.3, linestyle=':')
plt.show()
plt.close(fig)
Given a sine-wave of the form x = a sin(b(t+c)) + d, the period of the sine-wave is obtained as 2 * pi / b. Since b=1 (or by visual inspection), the period of our sine wave is 2 * pi. I can check the results obtained from other methods against this baseline.
Attempt 1: Autocorrelation
As I understand it (please correct me if I'm wrong), correlation can be used to see if one signal is a time-lagged copy of another signal (similar to how cosine and sine differ by a phase difference). So autocorrelation is testing a signal against itself to measure the times at which the time-lag repeats said signal. Using the example posted here:
result = np.correlate(x, x, mode='full')
Since x and t each consist of 100 elements and result consists of 199 elements, I am not sure why I should arbitrarily select the last 100 elements.
print("\n autocorrelation (shape={}):\n{}\n".format(result.shape, result))
autocorrelation (shape=(199,)):
[ 0.00000000e+00 -3.82130761e-16 -9.73648712e-02 -3.70014208e-01
-8.59889695e-01 -1.56185995e+00 -2.41986054e+00 -3.33109112e+00
-4.15799070e+00 -4.74662427e+00 -4.94918053e+00 -4.64762251e+00
-3.77524157e+00 -2.33298717e+00 -3.97976240e-01 1.87752669e+00
4.27722402e+00 6.54129270e+00 8.39434617e+00 9.57785701e+00
9.88331103e+00 9.18204933e+00 7.44791758e+00 4.76948221e+00
1.34963425e+00 -2.50822289e+00 -6.42666652e+00 -9.99116299e+00
-1.27937834e+01 -1.44791297e+01 -1.47873668e+01 -1.35893098e+01
-1.09091510e+01 -6.93157447e+00 -1.99159756e+00 3.45267493e+00
8.86228186e+00 1.36707567e+01 1.73433176e+01 1.94357232e+01
1.96463736e+01 1.78556800e+01 1.41478477e+01 8.81191526e+00
2.32100171e+00 -4.70897483e+00 -1.15775811e+01 -1.75696560e+01
-2.20296487e+01 -2.44327920e+01 -2.44454330e+01 -2.19677060e+01
-1.71533510e+01 -1.04037163e+01 -2.33560966e+00 6.27458308e+00
1.45655029e+01 2.16769872e+01 2.68391837e+01 2.94553896e+01
2.91697473e+01 2.59122266e+01 1.99154591e+01 1.17007613e+01
2.03381596e+00 -8.14633251e+00 -1.78184255e+01 -2.59814393e+01
-3.17580589e+01 -3.44884934e+01 -3.38046447e+01 -2.96763956e+01
-2.24244433e+01 -1.26974172e+01 -1.41464998e+00 1.03204331e+01
2.13281784e+01 3.04712823e+01 3.67721634e+01 3.95170295e+01
3.83356037e+01 3.32477037e+01 2.46710643e+01 1.33886439e+01
4.77778141e-01 -1.27924775e+01 -2.50860560e+01 -3.51343866e+01
-4.18671622e+01 -4.45258983e+01 -4.27482779e+01 -3.66140001e+01
-2.66465884e+01 -1.37700036e+01 7.76494745e-01 1.55574483e+01
2.90828312e+01 3.99582426e+01 4.70285203e+01 4.95000000e+01
4.70285203e+01 3.99582426e+01 2.90828312e+01 1.55574483e+01
7.76494745e-01 -1.37700036e+01 -2.66465884e+01 -3.66140001e+01
-4.27482779e+01 -4.45258983e+01 -4.18671622e+01 -3.51343866e+01
-2.50860560e+01 -1.27924775e+01 4.77778141e-01 1.33886439e+01
2.46710643e+01 3.32477037e+01 3.83356037e+01 3.95170295e+01
3.67721634e+01 3.04712823e+01 2.13281784e+01 1.03204331e+01
-1.41464998e+00 -1.26974172e+01 -2.24244433e+01 -2.96763956e+01
-3.38046447e+01 -3.44884934e+01 -3.17580589e+01 -2.59814393e+01
-1.78184255e+01 -8.14633251e+00 2.03381596e+00 1.17007613e+01
1.99154591e+01 2.59122266e+01 2.91697473e+01 2.94553896e+01
2.68391837e+01 2.16769872e+01 1.45655029e+01 6.27458308e+00
-2.33560966e+00 -1.04037163e+01 -1.71533510e+01 -2.19677060e+01
-2.44454330e+01 -2.44327920e+01 -2.20296487e+01 -1.75696560e+01
-1.15775811e+01 -4.70897483e+00 2.32100171e+00 8.81191526e+00
1.41478477e+01 1.78556800e+01 1.96463736e+01 1.94357232e+01
1.73433176e+01 1.36707567e+01 8.86228186e+00 3.45267493e+00
-1.99159756e+00 -6.93157447e+00 -1.09091510e+01 -1.35893098e+01
-1.47873668e+01 -1.44791297e+01 -1.27937834e+01 -9.99116299e+00
-6.42666652e+00 -2.50822289e+00 1.34963425e+00 4.76948221e+00
7.44791758e+00 9.18204933e+00 9.88331103e+00 9.57785701e+00
8.39434617e+00 6.54129270e+00 4.27722402e+00 1.87752669e+00
-3.97976240e-01 -2.33298717e+00 -3.77524157e+00 -4.64762251e+00
-4.94918053e+00 -4.74662427e+00 -4.15799070e+00 -3.33109112e+00
-2.41986054e+00 -1.56185995e+00 -8.59889695e-01 -3.70014208e-01
-9.73648712e-02 -3.82130761e-16 0.00000000e+00]
Attempt 2: Fourier
Since I am not sure where to go from the last attempt, I sought a new attempt. To my understanding, Fourier analysis basically shifts a signal from/to the time-domain (x(t) vs t) to/from the frequency domain (x(t) vs f=1/t); the signal in frequency-space should appear as a sinusoidal wave that dampens over time. The period is obtained from the most observed frequency since this is the location of the peak of the distribution of frequencies.
Since my values are all real-valued, applying the Fourier transform should mean my output values are all complex-valued. I wouldn't think this is a problem, except for the fact that scipy has methods for real-values. I do not fully understand the differences between all of the different scipy methods. That makes following the algorithm proposed in this posted solution hard for me to follow (ie, how/why is the threshold value picked?).
omega = np.fft.fft(x)
freq = np.fft.fftfreq(x.size, 1)
threshold = 0
idx = np.where(abs(omega)>threshold)[0][-1]
max_f = abs(freq[idx])
print(max_f)
This outputs 0.01, meaning the period is 1/0.01 = 100. This doesn't make sense either.
Attempt 3: Power Spectral Density
According to the scipy docs, I should be able to estimate the power spectral density (psd) of the signal using a periodogram (which, according to wikipedia, is the fourier transform of the autocorrelation function). By selecting the dominant frequency fmax at which the signal peaks, the period of the signal can be obtained as 1 / fmax.
freq, pdensity = signal.periodogram(x)
fig, ax = plt.subplots()
ax.plot(freq, pdensity, color='r')
ax.grid(color='k', alpha=0.3, linestyle=':')
plt.show()
plt.close(fig)
The periodogram shown below peaks at 49.076... at a frequency of fmax = 0.05. So, period = 1/fmax = 20. This doesn't make sense to me. I have a feeling it has something to do with the sampling rate, but don't know enough to confirm or progress further.
I realize I am missing some fundamental gaps in understanding how these things work. There are a lot of resources online, but it's hard to find this needle in the haystack. Can someone help me learn more about this?

Let's first look at your signal (I've added endpoint=False to make the division even):
t = np.linspace(0, 10*np.pi, 100, endpoint=False)
x = np.sin(t)
Let's divide out the radians (essentially by taking t /= 2*np.pi) and create the same signal by relating to frequencies:
fs = 20 # Sampling rate of 100/5 = 20 (e.g. Hz)
f = 1 # Signal frequency of 1 (e.g. Hz)
t = np.linspace(0, 5, 5*fs, endpoint=False)
x = np.sin(2*np.pi*f*t)
This makes it more salient that f/fs == 1/20 == 0.05 (i.e. the periodicity of the signal is exactly 20 samples). Frequencies in a digital signal always relate to its sampling rate, as you have already guessed. Note that the actual signal is exactly the same no matter what the values of f and fs are, as long as their ratio is the same:
fs = 1 # Natural units
f = 0.05
t = np.linspace(0, 100, 100*fs, endpoint=False)
x = np.sin(2*np.pi*f*t)
In the following I'll use these natural units (fs = 1). The only difference will be in t and hence the generated frequency axes.
Autocorrelation
Your understanding of what the autocorrelation function does is correct. It detects the correlation of a signal with a time-lagged version of itself. It does this by sliding the signal over itself as seen in the right column here (from Wikipedia):
Note that as both inputs to the correlation function are the same, the resulting signal is necessarily symmetric. That is why the output of np.correlate is usually sliced from the middle:
acf = np.correlate(x, x, 'full')[-len(x):]
Now index 0 corresponds to 0 delay between the two copies of the signal.
Next you'll want to find the index or delay that presents the largest correlation. Due to the shrinking overlap this will by default also be index 0, so the following won't work:
acf.argmax() # Always returns 0
Instead I recommend to find the largest peak instead, where a peak is defined to be any index with a larger value than both its direct neighbours:
inflection = np.diff(np.sign(np.diff(acf))) # Find the second-order differences
peaks = (inflection < 0).nonzero()[0] + 1 # Find where they are negative
delay = peaks[acf[peaks].argmax()] # Of those, find the index with the maximum value
Now delay == 20, which tells you that the signal has a frequency of 1/20 of its sampling rate:
signal_freq = fs/delay # Gives 0.05
Fourier transform
You used the following to calculate the FFT:
omega = np.fft.fft(x)
freq = np.fft.fftfreq(x.size, 1)
Thhese functions re designed for complex-valued signals. They will work for real-valued signals, but you'll get a symmetric output as the negative frequency components will be identical to the positive frequency components. NumPy provides separate functions for real-valued signals:
ft = np.fft.rfft(x)
freqs = np.fft.rfftfreq(len(x), t[1]-t[0]) # Get frequency axis from the time axis
mags = abs(ft) # We don't care about the phase information here
Let's have a look:
plt.plot(freqs, mags)
plt.show()
Note two things: the peak is at frequency 0.05, and the maximum frequency on the axis is 0.5 (the Nyquist frequency, which is exactly half the sampling rate). If we had picked fs = 20, this would be 10.
Now let's find the maximum. The thresholding method you have tried can work, but the target frequency bin is selected blindly and so this method would suffer in the presence of other signals. We could just select the maximum value:
signal_freq = freqs[mags.argmax()] # Gives 0.05
However, this would fail if, e.g., we have a large DC offset (and hence a large component in index 0). In that case we could just select the highest peak again, to make it more robust:
inflection = np.diff(np.sign(np.diff(mags)))
peaks = (inflection < 0).nonzero()[0] + 1
peak = peaks[mags[peaks].argmax()]
signal_freq = freqs[peak] # Gives 0.05
If we had picked fs = 20, this would have given signal_freq == 1.0 due to the different time axis from which the frequency axis was generated.
Periodogram
The method here is essentially the same. The autocorrelation function of x has the same time axis and period as x, so we can use the FFT as above to find the signal frequency:
pdg = np.fft.rfft(acf)
freqs = np.fft.rfftfreq(len(x), t[1]-t[0])
plt.plot(freqs, abs(pdg))
plt.show()
This curve obviously has slightly different characteristics from the direct FFT on x, but the main takeaways are the same: the frequency axis ranges from 0 to 0.5*fs, and we find a peak at the same signal frequency as before: freqs[abs(pdg).argmax()] == 0.05.
Edit:
To measure the actual periodicity of np.sin, we can just use the "angle axis" that we passed to np.sin instead of the time axis when generating the frequency axis:
freqs = np.fft.rfftfreq(len(x), 2*np.pi*f*(t[1]-t[0]))
rad_period = 1/freqs[mags.argmax()] # 6.283185307179586
Though that seems pointless, right? We pass in 2*np.pi and we get 2*np.pi. However, we can do the same with any regular time axis, without presupposing pi at any point:
fs = 10
t = np.arange(1000)/fs
x = np.sin(t)
rad_period = 1/np.fft.rfftfreq(len(x), 1/fs)[abs(np.fft.rfft(x)).argmax()] # 6.25
Naturally, the true value now lies in between two bins. That's where interpolation comes in and the associated need to choose a suitable window function.

Unexpected valley in Fourier Transformation using Scipy FFT

I am trying to perform Scipy FFT on my dataset. Basically, I have the acceleration in the time domain(obtained numerically) and I am just trying to perform Fourier transform, to obtain the spectrum. I have a theoretical expression for Fourier transformed acceleration in a small and large frequency limit. For large frequency, the Fourier transformed acceleration should drop exponentially. However, I am getting a valley in the graph after an initial decay. Below are my code and graph
a_w = []
for k in range(len(b)): # b is paramter to be varied
window = signal.kaiser(N, 30) # I am not sure about using Kaiser wind
ft = fft(solaccarr[k]*window)
ft = np.abs(ft[:N // 2])*1/N
freq = fftfreq(N, T)[:N // 2]
xf = np.linspace(0.0, 1.0 / (2.0 * float(T)), N // 2)
a_w.append(ft)
I am plotting the graph on a log-log scale. My question is that whether it's possible to get rid of the kinks in graph by appropriate use of windows or any other techniques?
Here is the dataset which I used

These valleys may correspond to the end of the main lobe of the Kaiser window after translations.
If the input signal features a finite number of well-defined frequencies (ex: a sum of two sine waves), its Fourier transform is a Dirac comb. Multiplying the signal and the window corresponds to convolving the DFT transform of the signal by the DFT transform of the window. As convolving with a Dirac signal corresponds to a translation, the outcome of the process is a finite sum of translated DFT transforms of the window.
The transform of the Kaiser window features a main lobe and side lobes separated by such valleys. Hence, the outcome may also features translated valleys. It can be tested by modifying 30 in window = signal.kaiser(N, 30): could you try figures like 0, 5, 6 and 8.6? It should translate the valley from left to right or right to left, as it modifies the width of the main lobe.
If you just want to get rid of the deep valleys, you can explore the exponential window eventually combined to an Hann window to form a Hann–Poisson window. This window does not feature any side lobes.
Finally, if your signal is periodic and if the length of the frame is a multiple of the period, there is no need for a window!

How to get frequency axis from an fft function?

So, I am probably missing something obvious, but I have searched through lots of tutorials and documentation and can't seem to find a straight answer. How do you find the frequency axis of a function that you performed an fft on in Python(specifically the fft in the scipy library)?
I am trying to get a raw EMG signal, perform a bandpass filter on it, and then perform an fft to see the remaining frequency components. However, I am not sure how to find an accurate x component list. The specific signal I am working on currently was sampled at 1000 Hz and has 5378 samples.
Is it just creating a linear x starting from 0 and going to the length of the fft'd data? I see a lot of people creating a linspace from 0 to sample points times the sample spacing. But what would be my sample spacing in this case? Would it just be samples/sampling rate? Or is it something else completely?

Here is an example.
First create a sine wave with sampling interval pre-determined. we will combine two sine waves with frequencies 20 and 40. Remember high frequencies might be aliased if the time interval is large.
#Import the necessary packages
from scipy import fftpack
import matplotlib.pyplot as plt
import numpy as np
# sampling freq in herts 20Hz, and 40Hz
freq_sampling1 = 10
freq_sampling2 = 20
amplitude1 = 2 # amplitude of first sine wave
amplitude2 = 4 # amplitude of second sine wave
time = np.linspace(0, 6, 500, endpoint=True) # time range with total samples of 500 from 0 to 6 with time interval equals 6/500
y = amplitude1*np.sin(2*np.pi*freq_sampling1*time) + amplitude2*np.sin(2*np.pi*freq_sampling2*time)
plt.figure(figsize=(10, 4))
plt.plot(time,y, 'k', lw=0.8)
plt.xlim(0,6)
plt.show()
Notice in the figure that two sine waves are superimposed. One with freq. 10 and amplitude 2 and the other with freq. 20 and amplitude 4.
# apply fft function
yf = fftpack.fft(y, time.size)
amp = np.abs(yf) # get amplitude spectrum
freq = np.linspace(0.0, 1.0/(2.0*(6/500)), time.size//2) # get freq axis
# plot the amp spectrum
plt.figure(figsize=(10,6))
plt.plot(freq, (2/amp.size)*amp[0:amp.size//2])
plt.show()
Notice in the amplitude spectrum the two frequencies are recovered while amplitude is zero at other frequencies. the Amplitude values are also 2 and 4 respectively.
you can use instead fftpack.fftfreq to obtain frequency axis as suggested by tom10
Therefore, the code changes to
yf = fftpack.fft(y, time.size)
amp = np.abs(yf) # get amplitude spectrum
freq = fftpack.fftfreq(time.size, 6/500)
plt.figure(figsize=(10,6))
plt.plot(freq[0:freq.size//2], (2/amp.size)*amp[0:amp.size//2])
plt.show()
We are only plotting the positive part of the amplitude spectrum [0:amp.size//2]

Once you feed your window of samples into the FFT call it will return an array of imaginary points ... the freqency separation between each element of returned array is determined by
freq_resolution = sampling_freq / number_of_samples
the 0th element is your DC offset which will be zero if your input curve is balanced straddling the zero crossing point ... so in your case
freq_resolution = 1000 / 5378
In general, for efficiency, you will want to feed an even power of 2 number of samples into your FFT call, important if you are say sliding your window of samples forward in time and repeatedly calling FFT on each window
To calculate the magnitude of a frequency in a given freq_bin (an element of the returned imaginary array)
X = A + jB
A on real axis
B on imag axis
for above formula its
mag = 2.0 * math.Sqrt(A*A+B*B) / number_of_samples
phase == arctan( B / A )
you iterate across each element up to the Nyquist limit which is why you double above magnitude
So yes its a linear increment with same frequency spacing between each freq_bin

Normalizing FFT spectrum magnitude to 0dB

I'm using FFT to extract the amplitude of each frequency components from an audio file. Actually, there is already a function called Plot Spectrum in Audacity that can help to solve the problem. Taking this example audio file which is composed of 3kHz sine and 6kHz sine, the spectrum result is like the following picture. You can see peaks are at 3KHz and 6kHz, no extra frequency.
Now I need to implement the same function and plot the similar result in Python. I'm close to the Audacity result with the help of rfft but I still have problems to solve after getting this result.
What's physical meaning of the amplitude in the second picture?
How to normalize the amplitude to 0dB like the one in Audacity?
Why do the frequency over 6kHz have such high amplitude (≥90)? Can I scale those frequency to relative low level?
Related code:
import numpy as np
from pylab import plot, show
from scipy.io import wavfile
sample_rate, x = wavfile.read('sine3k6k.wav')
fs = 44100.0
rfft = np.abs(np.fft.rfft(x))
p = 20*np.log10(rfft)
f = np.linspace(0, fs/2, len(p))
plot(f, p)
show()
Update
I multiplied Hanning window with the whole length signal (is that correct?) and get this. Most of the amplitude of skirts are below 40.
And scale the y-axis to decibel as #Mateen Ulhaq said. The result is more close to the Audacity one. Can I treat the amplitude below -90dB so low that it can be ignored?
Updated code:
fs, x = wavfile.read('input/sine3k6k.wav')
x = x * np.hanning(len(x))
rfft = np.abs(np.fft.rfft(x))
rfft_max = max(rfft)
p = 20*np.log10(rfft/rfft_max)
f = np.linspace(0, fs/2, len(p))
About the bounty
With the code in the update above, I can measure the frequency components in decibel. The highest possible value will be 0dB. But the method only works for a specific audio file because it uses rfft_max of this audio. I want to measure the frequency components of multiple audio files in one standard rule just like Audacity does.
I also started a discussion in Audacity forum, but I was still not clear how to implement my purpose.

After doing some reverse engineering on Audacity source code here some answers. First, they use Welch algorithm for estimating PSD. In short, it splits signal to overlapped segments, apply some window function, applies FFT and averages the result. Mostly as This helps to get better results when noise is present. Anyway, after extracting the necessary parameters here is the solution that approximates Audacity's spectrogram:
import numpy as np
from scipy.io import wavfile
from scipy import signal
from matplotlib import pyplot as plt
segment_size = 512
fs, x = wavfile.read('sine3k6k.wav')
x = x / 32768.0 # scale signal to [-1.0 .. 1.0]
noverlap = segment_size / 2
f, Pxx = signal.welch(x, # signal
fs=fs, # sample rate
nperseg=segment_size, # segment size
window='hanning', # window type to use
nfft=segment_size, # num. of samples in FFT
detrend=False, # remove DC part
scaling='spectrum', # return power spectrum [V^2]
noverlap=noverlap) # overlap between segments
# set 0 dB to energy of sine wave with maximum amplitude
ref = (1/np.sqrt(2)**2) # simply 0.5 ;)
p = 10 * np.log10(Pxx/ref)
fill_to = -150 * (np.ones_like(p)) # anything below -150dB is irrelevant
plt.fill_between(f, p, fill_to )
plt.xlim([f[2], f[-1]])
plt.ylim([-90, 6])
# plt.xscale('log') # uncomment if you want log scale on x-axis
plt.xlabel('f, Hz')
plt.ylabel('Power spectrum, dB')
plt.grid(True)
plt.show()
Some necessary explanations on parameters:
wave file is read as 16-bit PCM, in order to be compatible with Audacity it should be scaled to be |A|<1.0
segment_size is corresponding to Size in Audacity's GUI.
default window type is 'Hanning', you can change it if you want.
overlap is segment_size/2 as in Audacity code.
output window is framed to follow Audacity style. They throw away first low frequency bins and cut everything below -90dB
What's physical meaning of the amplitude in the second picture?
It is basically amount of energy in the frequency bin.
How to normalize the amplitude to 0dB like the one in Audacity?
You need choose some reference point. Graphs in decibels are always relevant to something. When you select maximum energy bin as a reference, your 0db point is the maximum energy (obviously). It is acceptable to set as a reference energy of the sine wave with maximum amplitude. See ref variable. Power in sinusoidal signal is simply squared RMS, and to get RMS, you just need to divide amplitude by sqrt(2). So the scaling factor is simply 0.5. Please note that factor before log10 is 10 and not 20, this is because we are dealing with power of signal and not amplitude.
Can I treat the amplitude below -90dB so low that it can be ignored?
Yes, anything below -40dB is usually considered negligeble

Using numpy/scipy to identify slope changes in digital signals?

I am trying to come up with a generalised way in Python to identify pitch rotations occurring during a set of planned spacecraft manoeuvres. You could think of it as a particular case of a shift detection problem.
Let's consider the solar_elevation_angle variable in my set of measurements, identifying the elevation angle of the sun measured from the spacecraft's instrument. For those who might want to play with the data, I saved the solar_elevation_angle.txt file here.
import numpy as np
import matplotlib.pyplot as plt
from matplotlib import gridspec
from scipy.signal import argrelmax
from scipy.ndimage.filters import gaussian_filter1d
solar_elevation_angle = np.loadtxt("solar_elevation_angle.txt", dtype=np.float32)
fig, ax = plt.subplots()
ax.set_title('Solar elevation angle')
ax.set_xlabel('Scanline')
ax.set_ylabel('Solar elevation angle [deg]')
ax.plot(solar_elevation_angle)
plt.show()
The scanline is my time dimension. The four points where the slope changes identify the spacecraft pitch rotations.
As you can see, the solar elevation angle evolution outside the spacecraft manoeuvres regions is pretty much linear as a function of time, and that should always be the case for this particular spacecraft (except for major failures).
Note that during each spacecraft manoeuvre, the slope change is obviously continuous, although discretised in my set of angle values. That means: for each manoeuvre, it does not really make sense to try to locate a single scanline where a manoeuvre has taken place. My goal is rather to identify, for each manoeuvre, a "representative" scanline in the range of scanlines defining the interval of time where the manoeuvre occurred (e.g. middle value, or left boundary).
Once I get a set of "representative" scanline indexes where all manoeuvres have taken place, I could then use those indexes for rough estimations of manoeuvres durations, or to automatically place labels on the plot.
My solution so far has been to:
Compute the 2nd derivative of the solar elevation angle using
np.gradient.
Compute absolute value and clipping of resulting
curve. The clipping is necessary because of what I assume to be
discretisation noise in the linear segments, which would then severely affect the identification of the "real" local maxima in point 4.
Apply smoothing to the resulting curve, to get rid of multiple peaks. I'm using scipy's 1d gaussian filter with a trial-and-error sigma value for that.
Identify local maxima.
Here's my code:
fig = plt.figure(figsize=(8,12))
gs = gridspec.GridSpec(5, 1)
ax0 = plt.subplot(gs[0])
ax0.set_title('Solar elevation angle')
ax0.plot(solar_elevation_angle)
solar_elevation_angle_1stdev = np.gradient(solar_elevation_angle)
ax1 = plt.subplot(gs[1])
ax1.set_title('1st derivative')
ax1.plot(solar_elevation_angle_1stdev)
solar_elevation_angle_2nddev = np.gradient(solar_elevation_angle_1stdev)
ax2 = plt.subplot(gs[2])
ax2.set_title('2nd derivative')
ax2.plot(solar_elevation_angle_2nddev)
solar_elevation_angle_2nddev_clipped = np.clip(np.abs(np.gradient(solar_elevation_angle_2nddev)), 0.0001, 2)
ax3 = plt.subplot(gs[3])
ax3.set_title('absolute value + clipping')
ax3.plot(solar_elevation_angle_2nddev_clipped)
smoothed_signal = gaussian_filter1d(solar_elevation_angle_2nddev_clipped, 20)
ax4 = plt.subplot(gs[4])
ax4.set_title('Smoothing applied')
ax4.plot(smoothed_signal)
plt.tight_layout()
plt.show()
I can then easily identify the local maxima by using scipy's argrelmax function:
max_idx = argrelmax(smoothed_signal)[0]
print(max_idx)
# [ 689 1019 2356 2685]
Which correctly identifies the scanline indexes I was looking for:
fig, ax = plt.subplots()
ax.set_title('Solar elevation angle')
ax.set_xlabel('Scanline')
ax.set_ylabel('Solar elevation angle [deg]')
ax.plot(solar_elevation_angle)
ax.scatter(max_idx, solar_elevation_angle[max_idx], marker='x', color='red')
plt.show()
My question is: Is there a better way to approach this problem?
I find that having to manually specify the clipping threshold values to get rid of the noise and the sigma in the gaussian filter weakens this approach considerably, preventing it to be applied to other similar cases.

First improvement would be to use a Savitzky-Golay filter to find the derivative in a less noisy way. For example, it can fit a parabola (in the sense of least squares) to each data slice of certain size, and then take the second derivative of that parabola. The result is much nicer than just taking 2nd order difference with gradient. Here it is with window size 101:
savgol_filter(solar_elevation_angle, window_length=window, polyorder=2, deriv=2)
Second, instead of looking for points of maximum with argrelmax it is better to look for places where the second derivative is large; for example, at least half its maximal size. This will of course return many indexes, but we can then look at the gaps between those indexes to identify where each peak begins and ends. The midpoint of the peak is then easily found.
Here is the complete code. The only parameter is window size, which is set to 101. The approach is robust; the size 21 or 201 gives essentially the same outcome (it must be odd).
from scipy.signal import savgol_filter
window = 101
der2 = savgol_filter(solar_elevation_angle, window_length=window, polyorder=2, deriv=2)
max_der2 = np.max(np.abs(der2))
large = np.where(np.abs(der2) > max_der2/2)[0]
gaps = np.diff(large) > window
begins = np.insert(large[1:][gaps], 0, large[0])
ends = np.append(large[:-1][gaps], large[-1])
changes = ((begins+ends)/2).astype(np.int)
plt.plot(solar_elevation_angle)
plt.plot(changes, solar_elevation_angle[changes], 'ro')
plt.show()
The fuss with insert and append is because the first index with large derivative should qualify as "peak begins" and the last such index should qualify as "peak ends", even though they don't have a suitable gap next to them (the gap is infinite).
Piecewise linear fit
This is an alternative (not necessarily better) approach, which does not use derivatives: fit a smoothing spline of degree 1 (i.e., a piecewise linear curve), and notice where its knots are.
First, normalize the data (which I call y instead of solar_elevation_angle) to have standard deviation 1.
y /= np.std(y)
The first step is to build a piecewise linear curve that deviates from the data by at most the given threshold, arbitrarily set to 0.1 (no units here because y was normalized). This is done by calling UnivariateSpline repeatedly, starting with a large smoothing parameter and gradually reducing it until the curve fits. (Unfortunately, one can't simply pass in the desired uniform error bound).
from scipy.interpolate import UnivariateSpline
threshold = 0.1
m = y.size
x = np.arange(m)
s = m
max_error = 1
while max_error > threshold:
spl = UnivariateSpline(x, y, k=1, s=s)
interp_y = spl(x)
max_error = np.max(np.abs(interp_y - y))
s /= 2
knots = spl.get_knots()
values = spl(knots)
So far we found the knots, and noted the values of the spline at those knots. But not all of these knots are really important. To test the importance of each knot, I remove it and interpolate without it. If the new interpolant is substantially different from the old (doubling the error), the knot is considered important and is added to the list of found slope changes.
ts = knots.size
idx = np.arange(ts)
changes = []
for j in range(1, ts-1):
spl = UnivariateSpline(knots[idx != j], values[idx != j], k=1, s=0)
if np.max(np.abs(spl(x) - interp_y)) > 2*threshold:
changes.append(knots[j])
plt.plot(y)
plt.plot(changes, y[np.array(changes, dtype=int)], 'ro')
plt.show()
Ideally, one would fit piecewise linear functions to given data, increasing the number of knots until adding one more does not bring "substantial" improvement. The above is a crude approximation of that with SciPy tools, but far from best possible. I don't know of any off-the-shelf piecewise linear model selection tool in Python.