基于磁共振电阻层析成像的图像重建算法研究

发布时间：2018-04-15 01:38

本文选题：磁共振电阻层析成像 + 反问题　；参考：《中国石油大学(华东)》2014年硕士论文

【摘要】：磁共振电阻层析成像(Magnetic Resonance Electrical Resistance Tomography,简称MRERT)技术是一种结合了传统电阻层析成像(Electrical Resistance Tomography,ERT)和磁共振成像技术(MRI)的新型非侵入式成像技术。该技术的显著优点是用MRI检测在被测对象边界注入电流的影响下,多相流型横截面的内部信息分布。MRI检测可以被用来重建图像化的横截面物理分布,且其分辨率较之ERT测量技术获得的图像有很大改进,这是由于MRERT测量技术的信息量显著增加。同时MRERT技术也减轻了ERT由于缺乏独立测量信息而导致的重建矩阵病态性。目前,MRERT技术已受到世界上生物医学、环境物理学和工业监测等领域的广泛关注。在这篇论文中,回顾了ERT作为一种多相流检测的成像检测技术目前的发展现状,并对其存在的问题做了总结。然后对于MRERT检测技术,突出介绍了其比较于ERT技术获得的图像质量高这一重大技术优势。其次对国内外MRERT技术的主要研究及应用做了简要回顾。接着进入对MRERT系统开发的技术细节介绍,包括正问题和反问题求解,多相流和电磁场分布的有限元建模,电势和电流密度分布,每个像素上由注入电流产生的拉莫尔频率偏移值的定量计算等。本文中还提出了一种基于灵敏度矩阵的算法用于定性地重建二维空间的电阻率分布。与ERT测量技术中为了获得所需的图像分辨率,通常需要多个注入电流不同的是,本文构建的MRERT系统仅需要一次注入电流,就可以产生足够的信息来重建图像。通过一系列采用灵敏度算法的流型仿真,测量了噪声的影响,多相空间分布以及油滴间距离等,研究了MRERT系统性能。采用Tikhonov正则化法改善灵敏度矩阵算法的成像质量。仿真结果表明,MRERT技术可以显著提高目前所有可通过ERT技术获得的图像质量及分辨率。尽管取得的结果只是初步的,本文的工作为今后将MRERT技术发展于工业过程检测奠定了重要的基础。
[Abstract]:Magnetic Resonance Electrical Resistance tomography (MRERT) is a new non-invasive imaging technique, which combines traditional electrical Resistance tomography and magnetic resonance imaging.The obvious advantage of this technique is that the MRI can be used to detect the internal information distribution of the cross section of the multiphase flow pattern under the influence of the injection current of the object under the influence of the measured object boundary. The detection can be used to reconstruct the physical distribution of the cross-section of the image.The resolution is much better than that obtained by ERT technique, which is due to the increase of information in MRERT measurement technology.At the same time, MRERT technology also alleviates the ill-condition of ERT reconstruction matrix caused by the lack of independent measurement information.At present, MRERT technology has been widely concerned in biomedicine, environmental physics and industrial monitoring.In this paper, the current development of ERT as a multi-phase flow detection imaging detection technology is reviewed, and its existing problems are summarized.Then, for the MRERT detection technology, the advantage of high image quality compared with ERT technology is highlighted.Secondly, the main research and application of MRERT technology at home and abroad are briefly reviewed.Then we introduce the technical details of MRERT system development, including forward and inverse problem solving, finite element modeling of multiphase flow and electromagnetic field distribution, electric potential and current density distribution,Quantitative calculation of the ramer frequency offset generated by the injection current on each pixel.In this paper, an algorithm based on sensitivity matrix is proposed to qualitatively reconstruct the resistivity distribution in two dimensional space.In order to obtain the required image resolution in ERT measurement technology, multiple injection currents are usually required. The proposed MRERT system needs only one injection current to generate enough information to reconstruct the image.The performance of MRERT system is studied by a series of flow pattern simulation using sensitivity algorithm. The effect of noise, the distribution of multi-phase space and the distance between oil droplets are measured.The Tikhonov regularization method is used to improve the imaging quality of the sensitivity matrix algorithm.Simulation results show that MRERT can significantly improve the image quality and resolution of all available images obtained by ERT technology.Although the results obtained are only preliminary, the work of this paper lays an important foundation for the future development of MRERT technology in industrial process detection.
【学位授予单位】：中国石油大学(华东)
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R445.2

【相似文献】