Storing volumetric data

The management of volumetric Raman spectroscopic data in RamanSPy is guided through the ramanspy.SpectralVolume class.

See also

As the ramanspy.SpectralVolume class extends the ramanspy.SpectralContainer class, most of its functionality is inherited from this class. Hence, users are advised to first check the documentation of the ramanspy.Spectrum class and the Generic data container tutorial.

Below, we will inspect some of the main features the ramanspy.SpectralVolume class provides on top of those inherited through the ramanspy.SpectralContainer class.

import numpy as np
import ramanspy

We can define a spectrum by providing a relevant 4D intensity data array and the corresponding Raman wavenumber axis, just as we initialise ramanspy.SpectralContainer instances. As an example, we will create a 50x50x10 spectroscopic volume, each point of which contains a Raman spectrum containing 1500 spectral points.

spectral_data = np.random.rand(50, 50, 10, 1500)
spectral_axis = np.linspace(100, 3600, 1500)

raman_volume = ramanspy.SpectralVolume(spectral_data, spectral_axis)

Then, we can use all features of the ramanspy.SpectralContainer class as usual. For instance,

raman_volume.shape

(50, 50, 10)

Another way to create ramanspy.SpectralVolume instances is by stacking ramanspy.SpectralImages instances along the z-axis. When doing that, the spatial dimensions of the images and their spectral axes must match.

raman_images = [ramanspy.SpectralImage(np.random.rand(50, 50, 1500), spectral_axis) for _ in range(5)]
raman_volume = ramanspy.SpectralVolume.from_image_stack(raman_images)

raman_volume.shape

(50, 50, 5)

Once we have initialised a ramanspy.SpectralVolume instance, we can visualise spectral slices across it

raman_volume.plot(bands=[1500, 2500])

• 
• 

[<Axes3D: title={'center': 'Raman volume'}>, <Axes3D: title={'center': 'Raman volume'}>]

as well as access individual layers from it and plot them

raman_image = raman_volume.layer(3)

raman_image.plot(bands=1500)

<Axes: title={'center': 'Raman image'}>