Signal normalisation

import hyperspy.api as hs
import numpy as np

Default normalisation

Create two unique Signal1D for example

s = hs.data.luminescence_signal()

s_shifted = np.roll(s.data, 200)
s_shifted = hs.signals.Signal1D(s_shifted)

s1 = (s/0.8 + hs.signals.Signal1D(s_shifted/4)) + 500
s2 = s + hs.signals.Signal1D(s_shifted/2)

Plot un-normalised signals

hs.plot.plot_spectra([s1,s2])

<Axes: xlabel='Energy (eV)', ylabel='Intensity'>

Plot both signals with default normalisation, of min and max intensity values to 0 and 1

hs.plot.plot_spectra([s1, s2], normalise = True)

<Axes: xlabel='Energy (eV)', ylabel='Normalised Intensity'>

Normalising to signal range

In the case that we want to normalise our signals to the background noise, we can create a custom normalisation function. The mean intensity over the index range 5 - 100 is specified to normalise the signals to

def normalise_mean_range(signal):
    data = signal.data

    normalise_range = signal.isig[5:100].data

    scale_factor = 1 / normalise_range.mean()

    return data * scale_factor

Plot signals with custom normalisation function passed as parameter

hs.plot.plot_spectra([s1,s2], normalise = normalise_mean_range)

<Axes: xlabel='Energy (eV)', ylabel='Normalised Intensity'>

Normalising to specified signal value

In the case that we want to normalise our signals to the second weaker peak in our spectra, we adjust our function to normalise the signals to the intensity at energy position 4.5 eV

def normalise_energy_value(signal):
    data = signal.data

    normalise_value = signal.isig[4.5].data

    scale = 1 / normalise_value

    return data * scale

Plot signals with custom normalisation function passed as parameter

hs.plot.plot_spectra([s1,s2], normalise = normalise_energy_value)

<Axes: xlabel='Energy (eV)', ylabel='Normalised Intensity'>