Application-specific optimization of regularization for electrical impedance tomography

Abstract

Single-step reconstruction of electrical impedance tomograms can have speed advantages. Key to obtaining good information from the tomograms is the regularization method used in reconstruction. For high-speed continuous monitoring of a patient or industrial process, there are important issues to consider when selecting a suitable value for the regularization parameter. A case study of a bubble column is presented. This demonstrates automation of the interpretation of sequences of tomograms and the optimal choice of regularization parameter.