Using outer boundary information for image reconstruction in magnetic induction tomography

Abstract

Over the last decade several experimental MIT systems have been developed which have overcome the difficulties of detecting low conductivity objects. There are several reports of systems that are capable of producing realistic 2D images of isolated conducting objects, or produces images showing the changes in electrical conductivity within an object using experimental data. However, absolute imaging of the conductivity distribution within an object presents a significant challenge. This paper focuses on the problem of reconstructing internal structure within low conductivity targets. One of the reasons this has proven to be a difficult task is that the flow of induced eddy currents inside the target is strongly dependent on the external boundary. Consequently, the boundary has a profound effect on the sensitivity of the mutual inductance values to changes in conductivity, which should be taken into account during image reconstruction. This typically involves time consuming eddy current calculations within an iterative procedure. This paper presents a method of producing sensitivity maps with lower computational cost than applying a forward problem solver which computes all the fields values every time. At the same time the technique introduced takes into account the boundary of the object. The maps obtained have been tested on real measurements. © Springer-Verlag 2007.