Imaging Range-ability of Particular ERT Reconstruction Algorithms

Yixin Ma1,2, Mi Wang1, Zhichu Zheng2, Yingxiang Wu2 and Nigel Holliday1

1 Centre for Particle and Colloid Engineering, University of Leeds, LS2 9JT, UK

2 The Institute of Mechanics, Chinese Academy of Sciences, Beijing 100080, P.R. China

ABSTRACT

The Electrical Resistance Tomography technique can produce color-labeled or gray images to show the conductivity distribution across a sensing domain. However, it has limited accuracy because of the difficulty in solving this type of non-linear electric field problem. This paper presents the results of several simulation studies to determine the spatial resolution and distinguish-ability, collectively terms as range-ability, of four ERT reconstruction algorithms. The algorithms are: the equi-potential line based back-projection (EBP) method, the sensitivity coefficient based back-projection (SBP) method, the generalized inverse matrix based iterative algorithm (GIMA) and the sensitivity theorem based iterative method (STM). By comparing the results, it was found that, the GIMA using prior-knowledge gives good range-ability for reconstructing distinct objects when the conductivity contrast between the objects and the background is small; the STM has the strongest ability of the four algorithms to recognize a continuous conductivity distribution and deliver the ‘true value’, the range-ability of the two back-projection algorithms are less affected by the conductivity contrast, however the object edges in the reconstructed image are far more blurred than that in STM and GIMA images. It was also noticed that the EBP method has better distinguish-ability than the SBP method for imaging distinct objects.

Keywords Electrical resistance tomography (ERT), Image reconstruction, Imaging range-ability