International Society for Industrial Process Tomography

Application of Electrical Impedance Tomography
for Monitoring Flood Embankments and Landfills

T. Rymarczyk^1,2*, G. Kłosowski³

¹ Research & Development Centre Netrix S.A., Wojciechowska 31, Lublin, Poland

² University of Economics and Innovation, Projektowa 4, Lublin, Poland

³ Lublin University of Technology, Nadbystrzycka 38A, Lublin, Poland

^*E-mail: tomasz@rymarczyk.com

ABSTRACT

The article presents the application concept of the EIT system for monitoring flood embankments and landfills. During the research, a comparison of selected seven regression models enabling the processing of input electrical signals into images, which included: Artificial Neural Networks (ANN), Fine Tree, Medium Tree, Coarse Tree, Linear Regression, Exponential Gaussian Process Regression (GPR) and Quadratic Support Vector Machine (SVM). The best results were obtained for the model based on artificial neural networks. The next part of the paper presents an innovative concept of a neural imaging system, the main feature of which is the use of a uniform vector of input signals for separate training of individual neural networks, each of which generates a single pixel of the output image. Thanks to this solution, the output image consisting of, for example, 17869 points requires the training of the same number of neural networks. The results of presented research confirm the high effectiveness of the described solution.

Keywords Electrical Impedance Tomography, Inverse Problem, Machine Learning, Regression Modes, Neural Networks.

Industrial Application: General