Research On Image Reconstruction Of Electrical Resistance Tomography Based On Kalman Filter

Electrical Resistance Tomography (ERT) is one of Process Tomographytechniques based on the sensitivity field principle of conductance. According to theelectrical properties of the interior structures, resistivity distribution is obtainedthrough the special design of the sensor electrode in non-intrusive or non-contactway. Then the image of the object field is reconstructed. ERT has the attractivedeveloping potential in the field of industrial testing application and clinicaldiagnosis application, for its low cost, fast response, harmlessness, andnon-destructive visible measurement techniques.The important meaning of ERT to the measurement developing is summarizedin this paper. It clarifies the basic principles of ERT, also analyzes the structure andtechnical characteristic. In addition, the current status and broad application prospectof ERT are discussed. It is emphasized that the main problems result in the difficultyof its practical application. Therefore, the calculation of sensitivity field and imagereconstruction algorithm are the main point studied in this paper. The main works arelisted as follows:First, this paper establishes the simulation of the circular field of2D ERT withthe software COMSOL and Matlab. The information of forward problem areobtained, including electric potential and the boundary voltage etc. in ERT system,to research the parameter selection of the sensor electrode and the incentive-measuremode, to establish the solving process of forward problem.Second, the calculation method of the sensitivity coefficient field of ERT basedon COMSOL Script software and the potential distribution is presented in this paper.These provide powerful data for image reconstruction.Finally, the theory of the inverse problem is analyzed. On the basis of the studyto LBP and Landweber algorithm, this paper proposes a fast image reconstructionalgorithm with Kalman filter. At the same time, the performance evaluationparameters of image reconstruction are introduced. The object of this paper is on the research of12-electrode ERT system. This paper sets simulation experiments for sixkinds of typical conductivity distribution. The six experiments prove that using theKalman filter algorithm has better correlation coefficient, smallererror, and the clearest image. The results show that the proposed algorithm onKalman filter has the characteristic of better precision of image reconstruction, andbetter robustness to noise.