A multiresolution reconstructive algorithm based on network theory for electrical capacitance tomography

Electrical capacitance tomography technique reconstructs dielectric constant distribution in an object by measuring the capacitances between the eletrode pairs which are mounted around this object. Because of the limitation of measurement condition, the measured data are imcomplet. This paper describes a multiresolution reconstructive algorithm which is based on network theory for electrical capacitance tomography technique. The dielectric constant distribution of flow of two components in a pipeline is reconstructed. The algorithm is as follows: Firstly, construct a rough, first level system model, and assume the dielectric constant distribution of the region to be reconstructed. After iteration, the dielectic constant of each unit can be reconstructed. Secondly, construct a finer, second level the system model and determine the initial dielectric constant of each unit in the region to be reconstructed according to related information between two levels. After iteration, the image of the pipeline's cross section can be reconstructed. The results of simulated experiments about different kinds of medium distributions show that this algorithm is effective and can converge. Supported by the National Natural Science Foundation of China Ma Ning: born in 1970, Ph. D.