Investigation Into Image Reconstruction Algorithm And Software Optimization Of ERT

Electrical Tomography (ET) has been rapidly developed since the 80th of last century when it was proposed. Due to the advantages of being non-intrusive and non- radiant, rapid response, simple structure and low cost, ET has broadly prospects in the area of clinical application and industrial oil-gas-water multi-phase flow measu- rement.Now two of the most mature Electrical Tomography technologies are Electrical Capacitance Tomography(ECT) and Electrical Resistance Tomography (ERT). ECT is suitable to measure the non-conducting medium as the continuous phase in the multi- phase flow, while ERT is suitable to measure the conductive medium.Although ET has been rapid developmented for decades, there are still many problems: Satisfactory images haven’t been achieved due to the ill-posedness of ET inverse problem. Solving the inverse problem needs better algorithm.And the ET software system has lagged behind leading to non-humanized design, difficult to operate and single function. To solve the problems, thorough research and improvements have been conducted mainly on image reconstruction algorithms and optimization of visualization software for ET system. The main work and results are as follows.1. Research and improvement on Image Reconstruction Algorithms for ERTIn the third chapter, principles of algorithms have been summed up and applied to ERT image reconstruction, which is commonly adopted for non-linear inverse problems. A symmetric conjugate gradient algorithm which is based on conjugate gradient algorithm was proposed, which had the faster convergent rate. Then the symmetric conjugate gradient algorithm based on QR decomposition was proposed. It is theoretically proven that this algorithm can converge only by one step, and it has good stability and real-time performance. Experimental results demonstrate that the one-step stable algorithm is of fine real-time performance and robustness.2. Optimization of visualization software TJU-ERT-V for industrial ERT SystemTo improve the performance of Electrical Resistance Tomography in the applica- tion of multi-phase flow, a real-time ERT software system named as TJU-ERT-V was designed. TJU-ERT-V is designed with aid of the Visual C++6.0 programming environment, in which MFC and OpenGL are applied. TJU-ERT-V not only can be used for on-line detecting, but also off-line analysis. This software system which contained four fast image reconstruction algorithms (Sensitivity Algorithm, Pre- iteration Method, Standard Tikhonov Algorithm and QR_CG_Method) can carry out the image reconstruction for 16 electrode ERT system. Also, it implements the functions of high-speed communication via USB, algorithm embedment, graphics and image processing, history data storage and playback, and the like.