State estimation in process tomography - Three-dimensional impedance imaging of moving fluids

摘要

In this paper, we consider three-dimensional impedance imaging of rapidly varying objects. We especially concentrate on the case where the target is a moving fluid and the objective is to track the concentration distribution of a substance dissolved in the fluid. The observations are made as in ordinary electrical impedance tomography (EIT), but in the reconstruction we employ the convection-diffusion model to yield information on the temporal behavior of the object. The observation model of EIT together with the evolution model constitute the state-space representation of the system, and the reconstruction problem of EIT can be described as a state estimation problem. The state estimation problem is solved using the Kalman filter and the fixed-interval smoother algorithms. The performance of the state estimation approach is evaluated in a simulation study. Copyright (c) 2007 John Wiley & Sons, Ltd.