ERT field data inversion with structural boundary constraints

geohang commented 3 years ago

Problem description

Hi, I am a geophysical student and learning inversion. I was running the ERT field data example from the transform2021 https://github.com/gimli-org/transform2021. The example is pretty good and easy to understand, especially the inversion example with a known structural boundary as prior information. The inversion results with a known structural boundary are impressive in that the structure is totally constrained. I am just curious about how the pygimli can achieve that. Is that similar to the work (Aukenand Christiansen, 2004)? If possible, would you mind providing the paper information about the impressive technology.

Thanks Hang

Auken, E. and Christiansen, A.V., 2004. Layered and laterally constrained 2D inversion of resistivity data. Geophysics, 69(3), pp.752-761.

Your environment

Operating system: e.g. Windows, Python version: 3.7 pyGIMLi version: 1.1.1+5.gc107727a Way of installation: e.g. Conda package

Steps to reproduce

geo = pg.meshtools.createParaMeshPLC(data, paraMaxCellSize=100)
line = pg.meshtools.createLine(start=[60, 0], end=[160, 0], marker=1)
geo += line
ax, _ = pg.show(geo);
ax.set_xlim(-20, 180);
ax.set_ylim(-30, 5);
mesh = pg.meshtools.createMesh(geo, quality=34)
ax, _ = pg.show(mesh)
ax.set_ylim(-30, 10)
ax.set_xlim(-20, 180);
mgrConstrained = ert.ERTManager()
mgrConstrained.invert(data=data, verbose=True, lam=50, mesh=mesh, maxIter=5)
mgrConstrained.showResult(cMin=1, cMax=500, xlabel="x (m)", ylabel="z (m)");

halbmy commented 3 years ago

The regularization with structural constraints has nothing to do with the works of the Aarhus group, but has been described in various publications:

Rücker (2011): Advanced Electrical Resistivity Modelling and Inversion using Unstructured Discretization
Günther, T., Musmann, P., Schaumann, G. & Grinat, M. (2011): Imaging of a fault zone by a large-scale dc re- sistivity experiment and seismic structural information. - Ext. abstr., 17th EAGE Near Surface, 12.-14.09.2010; Leicester.
Doetsch, J., Linde, N., Pessognelli, M., Green, A.G. & Günther, T. (2012).: Constraining 3-D electrical resistance tomography with GPR data for improved aquifer characterization. Journal of Applied Geophysics 78, 68-76, doi:10.1016/j.jappgeo.2011.04.008.
Jiang, C., Igel, J., Dlugosch, R., Müller-Petke, M., Günther, T., Helms, J., Lang, J. & Winsemann (2020): Magnetic resonance tomography constrained by ground-penetrating radar for improved hydrogeophysical characterisation, Geophysics 85(6), JM13-JM26, doi:10.1190/geo2020-0052.1.

geohang commented 3 years ago

Thank you very much!

Hang

gimli-org / gimli