基于正则化参数优化和边界聚类的电阻抗成像研究
Research on electrical impedance imaging based on regularization parameter optimization and boundary clustering
王苏煜 1戎舟 1袁晶晶1
作者信息
- 1. 南京邮电大学自动化学院人工智能学院 南京 210046
- 折叠
摘要
电阻抗成像是一种无损伤的功能成像技术,由于逆问题具有不适定性、不稳定性等特点,往往存在重构图像的分辨率不高、伪影较大等问题.将Tikhonov和全变量(TV)两种正则化算法的罚函数进行组合应用,提出将粒子群算法用于组合罚函数的正则化参数优化,把图像质量指标(artifact level,AL)作为粒子群算法的适应度值,从而确定最优正则化参数,通过牛顿迭代法获得电导率,为了进一步去除伪影,将Niblack算法与边界聚类算法相结合,对求得的电导率进行处理,得到最终的电导率分布.仿真和实测结果均表明,该方法重建的图像能够更加准确地反映电场内目标物体的位置信息,有效的抑制伪影,提高了重建效果.
Abstract
Electrical impedance tomography(EIT)is a non-destructive functional imaging technique,due to the inherent characteristics of the inverse problem,such as discomfort and instability,there are often problems such as low resolution and large artifacts of reconstructed images.In this paper,the penalty functions of two regularization algorithms Tikhonov and TV are combined,and particle swarm optimization is proposed to optimize the regularization parameters of the combined penalty function.The image quality indicator AL is taken as the fitness value of the particle swarm algorithm,so as to determine the optimal parameters,and the conductivity is obtained through iterative solution by Newton iteration method.In order to further remove artifacts,Niblack algorithm and boundary clustering algorithm are combined to process the obtained conductivity,and the final conductivity distribution is obtained.The results of simulation and measurement show that the reconstructed image can reflect the position information of the target object more accurately,effectively inhibit the generation of artifacts,and improve the reconstruction effect.
关键词
电阻抗成像/逆问题/Tikhonov正则化算法/粒子群算法/边界聚类算法/图像重建Key words
electrical impedance tomography/boundary clustering algorithm/Tikhonov regularization algorithm/particle swarm optimization algorithm/inverse problem/image reconstruction引用本文复制引用
出版年
2024
NETL
NSTL
万方数据