An image reconstruction algorithm for electrical capacitance tomography based on simulated annealing particle swarm optimization

作     者：Wang, P. Lin, J.S. Wang, M. 

作者机构：School of Electric and Control Xi'an University of Science and Technology Xi'an China Department of Computer Science and Information Engineering National Chin-Yi University of Technology Taichung Taiwan 

出 版 物：《Journal of Applied Research and Technology》 (J. Appl. Res. Technol.)

年 卷 期：2015年第13卷第2期

页      面：197-204页

学科分类：0810[工学-信息与通信工程] 12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 0805[工学-材料科学与工程（可授工学、理学学位）] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：National Natural Science Foundation of China, NSFC, (201314) Ministry of Education of the People's Republic of China, MOE, (2011508) Xi'an University of Science and Technology, XUST Scientific Research Foundation for Returned Scholars of Ministry of Education 

主　　题：Electrical capacitance tomography Least squares support vector machines Particle swarm optimization Simulated annealing algorithm 

摘      要：In this paper, we introduce a novel image reconstruction algorithm with Least Squares Support Vector Machines (LS SVM) and Simulated Annealing Particle Swarm Optimization (APSO), named SAP. This algorithm introduces simulated annealing ideas into Particle Swarm Optimization (PSO), which adopts cooling process functions to replace the inertia weight function and constructs the time variant inertia weight function featured in annealing mechanism. Meanwhile, it employs the APSO procedure to search for the optimized resolution of Electrical Capacitance Tomography (ECT) for image reconstruction. In order to overcome the soft field characteristics of ECT sensitivity field, some image samples with typical flow patterns are chosen for training with LS-SVM. Under the training procedure, the capacitance error caused by the soft field characteristics is predicted, and then is used to construct the fitness function of the particle swarm optimization on basis of the capacitance error. Experimental results demonstrated that the proposed SAP algorithm has a quick convergence rate. Moreover, the proposed SAP outperforms the classic Landweber algorithm and Newton-Raphson algorithm on image reconstruction. © 2015 Universidad Nacional Autónoma de México, Centro de Ciencias Aplicadas y Desarrollo Tecnológico. All Rights Reserved.