Research On Image Reconstruction Algorithm Based On Soft-field Characteristics For Electrical Resistance Tomography

As a kind of no radiation, no invasion, high speed of response, simple structure, low cost, wide application range, and fine safety measurement, Electrical Resistance Tomography (ERT) is a kind of process tomography technique based on the sensitive principle of capacitance. It has been the most popular research direction and the main development technique of process tomography with a wide application prospect in the field of industry. Currently it is focused on the stage of the laboratory research and there is a great distance to the large-scale industrialization.So it is very equivalent significance to obtain an intelligent software and simulation performance environment in which process of image reconstruction in ERT. The research in this thesis mainly includes:1. Simulation software is developed for the ERT system. It can provide functions in which you can easily set the circular pipe, resistance sensors and flow distribution parameters. With the data above, the positive and opposite problem of the image reconstruction algorithm can be researched conveniently for ERT.2. In accordance with the development trends and important research significance in industrial applications to the characteristics of multiphase flow system, it is pointed out that the superiority of ERT in the process tomography analysis of imaging non-destructive testing, non-invasive testing and other related aspects comparing to the current epidemic of several process tomography technology.3. The 12-electrode electrical resistance tomography system as the object to analysis the structure, considering the development of the current status of ERT, which is predicted its future prospects for the macro. 4. A new type of neural network image reconstruction algorithm applied to the process of resistance tomography image reconstruction is proposed. For the qualitative issues of electrical resistance tomography image reconstruction algorithm, it divided the entire distribution of sensitivity field into six sub-networks, and reduced the size of the original network. It improved the algorithm speed in training and the image quality, especially in the flow pattern.