International Society for Industrial Process Tomography

EIT Velocity Field Estimation via Pixel-to-Pixel Least-Squares Matching

Shengheng Liu, Jiabin Jia^*

School of Engineering, The University of Edinburgh, Edinburgh EH9 3JL, UK

^*Email: jiabin.jia@ed.ac.uk

ABSTRACT

Industrial process tomography generally aims to continuously interrogate the dynamic behaviour within the process vessels. In particular, there is a need to measure not only the location of inclusions but also their velocity of movement. Previously, we have enhanced the spatial resolution of electrical impedance tomography (EIT) modality by solving the inverse problem with a structure-aware sparse Bayesian learning (SA-SBL) based algorithm. On this basis, we present a simple and reliable approach based on pixel-to-pixel least-squares matching in this work to estimate and visualize the velocity field from the sequential EIT frames. Real-data experiment results demonstrate that the proposed approach can achieve an improved performance compared to existing methods, as high-quality EIT frames and robust pixel-to-pixel correlator both facilitate accurate velocity field estimation.

Keywords Process tomography, least-squares matching, sparse Bayesian learning, velocity field

Industrial Application General