New Reconstruction Algorithm for Electrical Capacitance Tomography

Thomas Loser1, Radoslaw Wajman2 and Dieter Mewes1

1Institute of Process Engineering, University of Hannover, Callinstr. 36, 30167 Hannover, Germany Email: dms@c36.uni-hannover.de, Website: www.ifv.uni-hannover.de

2Computer Engineering Department, Technical, University of Lodz, al. Politechniki 11, 90-924 Lodz, Poland, Website: www.kis.p.lodz.pl

ABSTRACT

For image reconstruction in electrical capacitance tomography (ECT) different implementations of the linear backprojection algorithm and iterative reconstruction techniques have become the standard in most applications. In these algorithms usually sensitivity matrices are used. Most implementations suffer from a smoothening effect and lack of accuracy due to their smooth sensitivity matrices and the so called "soft" field problem in ECT, respectively. Therefore it is usually difficult to reconstruct objects with sharp contours in capacitance tomography. In the present paper a novel reconstruction algorithm based on a Finite-Element-Method (FEM)-calculation is introduced. This algorithm is derived directly from linear X-ray-tomography, where local values can be reconstructed along the rays. Similar to that, the reconstruction in ECT is done along integral lines of the electric field. Weighting matrices are calculated along these integral lines. These weighting matrices are implemented within iterative reconstruction techniques like ART (Algebraic Reconstruction Technique). The new method can be implemented in many reconstruction techniques by using weighting matrices along the electric field lines instead of sensitivity matrices. With the new algorithm even objects with sharp contours can be reconstructed. The results are more detailed and accurate tomograms.

Keywords electrical capacitance tomography, image reconstruction, FEM, electric field lines, sensitivities, ART