Research On Image Reconstruction Algorithm For Electrical Capacitance Tomography

Electrical Capacitance Tomography, abbreviated as ECT, is a process imaging technology based on capacitance sensing mechanism. At present the theory of ECT technology is still not adequate, mutual restrict between imaging speed and image accuracy limit the wide application of the ECT technology. Based on the status of current development, research on image reconstruction algorithm can significantly improve the result of image reconstruction, which can promote the development of ECT technology.ECT image reconstruction is a discrete ill-posed inverse problem, the ill-posed and morbid character as well as its inherent "soft field" feature have serious effect on the results of image reconstruction. This thesis focuses on the research of image reconstruction from the perspective of improving reconstruction speed and realizing application of the algorithm. The main work are as follows:(1). ECT principles is deeply analyzed from the point of forward and inverse problem. Several tools are utilized including singular value decomposition and condition number to analyze the ill-conditioning performance of the discrete inverse problem and its effect on the stability of the problem solution.(2). Three representative image reconstruction algorithms as well as their respective limitations in ECT application are researched. Then the focus of this paper is determined as realizing simple and fast imaging.(3). Exploring how to improve the speed of Landweber iterative algorithms for image reconstruction. After the optimization of Landweber iterative process by appropriate mathematical transformation, single-step Landweber algorithm whose iterative matrix can be calculated off-line.is deduced without changing the principle of algorithm itself.(4). Researching on the difficulty in determining regularization parameter of Tikhonov algorithm for image reconstruction. At present the parameter is choosed through experience value or complex method. In this paper, the largest singular value method is proposed to determine the regularization parameter, which can effectively reduce the difficulty in regularization algorithm application.(5). The image reconstruction methods under different permittivity is studied through analysis of the sensitive field of sensor area. As for high dielectric constant, the "soft field" feature is quite obvious, therefore its impact on the image reconstruction can not be ignored. Finally, a new method to construct the sensitivity matrix is presented to solve the problem that conventional method can not be applied in high and unknown permittivity.(6). An8-electrode ECT experimental system is set up based on NI LabVIEW software.8-electrode sensor is fabricated, data acquisition circuit which was designed by predecessors is repaired and debugged so that it can meet the needs of this subject. Utilizing the system to test the reconstruction algorithm and explore an primary method based on ECT technology to detect the safety of common flammable liquids.