Application of Artificial Neural Network for the Inversion of Electrical Resistivity Data

O. L. Johnson, A. P. Aizebeokhai


The inversion of most geophysical data sets is complex due to the inherent non-linearity of the inverse problem. This usually leads to non-uniqueness of solutions to the inverse problem. Artificial neural network (ANN) has been used effectively to address several non-linear and non-stationary inverse problems. This study is essentially an assessment of the effectiveness of estimating subsurface resistivity model parameters from apparent resistivity measurements using ANN. Multi-layered earth models for different geologic environments were used to generate synthetic apparent resistivity data. The synthetic apparent resistivity data were generated using linear filter method embedded in the RES1D program. Neural network toolbox on MATLAB was used to design, train and test a developed neural network that was employed in the inversion of the apparent resistivity sounding data sets. Resilient feed-forward back propagation algorithm was used to train the network. The network was trained with 50% of the synthetic apparent resistivity data sets and their corresponding multi-layered earth models. 25% of the data set was used to test the network and the network was validated with another 25% of the data set. The network was then used to invert field data obtained from Iyanna-Iyesi, southwestern Nigeria. The results obtained from ANN responses were compared with that of a conventional geoelectrical resistivity inversion program (WINRESIST); the results indicate that ANN is effective in the inversion of geoelectrical resistivity sounding data for multi-layered earth models.


Artificial Neural Network (ANN)

Full Text:



A.P. Aizebeokhai, 2D and 3D geoelectrical resistivity imaging: theory and field design, Scientific Research and Essays 5 (23), 3592 – 2605.

H. Demuth, M. Beale and M. Hagan, Neural Network Toolbox, 1st edition, Natick, Mass, MathWorks 9 (4), 259 – 265.

O. Koefoed, Resistivity sounding on an earth model containing transition layers with linear change of resistivity with depth, Geophysical Prospecting 27 (4), 862 – 868.

L. Lines and S. Treitel, A review of least-squares inversion and its application to geophysical problems, Geophysical Prospecting 32 (2), 159 – 186.

G.F. Luger and W.A. Stubblefield. Artificial Intelligence: Structures and Strategies for Complex Problem Solving, 2nd edition, Benjamin/Cumming Publishing, Redwood City, California, p. 850 (1993).

J.L. McClelland, D.E. Rumelhart and G.E. Hinton, The Appeal of Parallel Distributed Processing, in Parallel Distributed Processing: Explorations in the Microstructure of Cognition Foundations, MIT Press, Cambridge, p. 34 (1986).

J. Stephen, C. Manoj and S. Singh, A direct inversion scheme for deep resistivity sounding data using artificial neural networks, Journal of Earth System Science 113 (1) (2004), 49 – 66.

A. Tarantola, Inverse Problem Theory and Methods for Model Parameter Estimation, 1st edition, Society for Industrial and Applied Mathematics, Philadelphia, PA, p. 358 (2005).


eISSN 0975-5748; pISSN 0974-875X