MNR图像重建算法中正则化因子研究.pdf

Computer Engineering and Applications 计算机工程与应用 2011 ，47 （21 ） 13 MNR 图像重建算法中正则化因子研究 1，2 1 肖理庆 ，王化祥 XIAO Liqing1，2 ，WANG Huaxiang1 1.天津大学 电气与自动化工程学院，天津 300072 2.徐州工程学院 信电工程学院，江苏 徐州 221111 1.School of Electrical Engineering Automation ，Tianjin University ，Tianjin 300072 ，China 2.School of Information Electrical Engineering ，Xuzhou Institute of Technology ，Xuzhou ，Jiangsu 221111 ，China XIAO Liqing ，WANG Huaxiang.Research on regularization factor of Modified-Newton-Raphson algorithm for image re- construction.Computer Engineering and Applications ，2011 ，47 （21 ）：13-16. Abstract ：Aiming to improve the precision of the image reconstruction algorithm when solving the inverse problem in elec- trical resistance tomography ，research on the regularization factor of modified-Newton-Raphson algorithm is carried out.Moti- vated by the idea of decreasing inertia weight used in the improved particle swarm optimization ，the upper bound of the reg- ularization factor is updated automatically according to the imaging quality during the iteration process.Hence a novel im- proved Newton-Raphson reconstruction algorithm is proposed and applied to image reconstruction of two-phase typical flow regimes—stratified flow ，bubble flow ，annular flow ，core flow and compound flow.Simulation results demonstrate that ，com- pared with iterative linear back projection algorithm and modified Newton-Raphson algorithm ，the new proposed algorithm can improve the imaging accuracy effectively under the same experimental condition. Key words ：electrical resistance tomography ；image reconstruction algorithm ；modified-Newton-Raphson algorithm ；regulariza- tion factor ；particle swarm optimization ；inertia weight 摘 要：为了提高电阻层析成像图像重建算法求解逆问题精度，对修正牛顿-拉夫逊算法中正则化因子进行了研究。借鉴改进粒 子群算法中惯性权重递减策略，根据算法迭代过程中成像精度，自动更新正则化因子的最大值，提出一种新的改进牛顿-拉夫逊 图像重建算法，应用于两相流典型流型——层状流、泡状流、环状流、中心流及复合流型图像重建。仿真实验结果表明，相同实验 条件下，相比