Introduce covariance function in the characterization module

[fussegli]

Introduce the covariance function used for calculation of the specific surface area and the particle mean diameter as done in previous work https://iopscience.iop.org/article/10.1149/07711.1095ecst.

That's low priority although, and will likely be added when i will transition the characterization module to app designer. Reason being I do not think calculating surface area from 3D imaging is a good idea due to the fractal behavior of the surface area (although still interesting to compare electrodes imaged at the same voxel size). Anyway, the difference between the direct method and the covariance function has been investigated in the work cited above for NMC electrodes and: "Sp(geometric covariogram) / Sp(direct method x2/3) ranges from 1.028 to 1.094 for the 16 volumes. It corresponds to a relative error (for the direct method) inferior to 9%." The method is interesting to get particle diameter although, but the issue with this method is that it requires a large field of view to get a smooth two-point correlation function and get the asymptot.