摘要
提出了一种基于贝叶斯理论的电容层析成像(electrical capacitance tomography,ECT)图像重建算法。与现有的ECT图像代数重建算法不同,该算法将介质分布的先验信息和噪声信息等统计信息引入到ECT图像重建中。文中针对5种典型介质分布进行了数值仿真实验,结果表明该算法对于两相流中不同的介质分布均具有良好的适应性。迭代步数和先验概率的协方差取值对图像重建结果有较大影响。在迭代过程中,该算法所需的增益矩阵的更新不依赖于输入,因而可以离线预先计算,从而减小图像重建的计算压力,提高该算法重建速度。
This paper proposes an iterative image reconstruction algorithm based on Bayesian theorem for electrical capacitance tomography （ECT）. In order to solve the inverse problem of ECT, statistical information, including the prior probability of the permittivity distribution and the noise inform- ation in the measurement data, is taken into account, which is different to the existing ECT image reconstruction algorithms. Simulations with regard to five typical permittivity distributions are investigated. The numerical results show the feasibility and efficiency of this iterative algorithm. The effects of iteration steps and the variance with the prior probability on the quality of image reconstruction are investigated. Since the gain matrix is independent on the input during the iteration, it is suggested that the gain matrix be calculated offiine in advance to accelerate the iteration procedure and improve the speed of the algorithm.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2008年第11期65-71,共7页
Proceedings of the CSEE
基金
国家自然科学基金项目(60532020
60204003)~~
关键词
电容层析成像
图像重建
贝叶斯理论
高斯分布
迭代算法
electrical reconstruction
Bayesian iterative algorithm capacitance tomography
image theorem
Gaussian distribution
iterative algorithm
