标题二号，黑体，英文字体为黑体，1-仪器仪表学报.docVIP

一种基于聚类电阻层析成像静态图像重建算法 肖理庆1 王化祥2 （1.天津大学精密仪器与光电子工程学院 天津 300072; 2.天津大学电气与自动化工程学院 天津 300072） 摘　要：为了提高静态图像重建算法求解电阻层析成像逆问题的实时性与精度,提出一种基于敏感场区域内三角形有限元聚类的静态图像重建算法.新算法初期采用改进牛顿-拉夫逊算法,并根据算法振荡次数超过允许最大值时所对应的最优图像重建结果,对敏感场区域内所有三角形有限元进行聚类,最后利用改进粒子群算法求解电阻层析成像逆问题,从而实现静态图像重建.新算法每隔一定迭代次数对敏感场区域内所有三角形有限元进行重新聚类,以减小粒子维数,提高新算法求解电阻层析成像逆问题的精度.实验结果表明,相比改进牛顿-拉夫逊算法及组合算法,新算法效果最理想. 关键词：静态图像重建;电阻层析成像;逆问题;实时性;聚类 中图分类号:TH701 文献标识码:A 国家标准学科分类代码:460.4010 Absolute Image Reconstruction Algorithm Based on Clustering for ERT XIAO Liqing1, WANG Huaxiang2 (1. School of Precision Instrument Optoelectronics Engineering, Tianjin University, Tianjin 300072, China; 2. School of Electrical Engineering Automation, Tianjin University, Tianjin 300072, China) Abstract：In order to improve the real-time and accuracy of absolute image reconstruction algorithm, a new method based on triangular finite elements clustering in the sensitive field of electrical resistance tomography was proposed. Firstly, the improved NR algorithm was adopted, and the triangular finite elements in the sensitive field were clustered according to the best result, when the oscillation frequency of the novel absolute algorithm exceeded the maximum allowed value, finally, the improved PSO algorithm was adopted to realize absolute image reconstruction. The triangular finite elements in the sensitive field were re-clustered after a certain number of iterations during the iteration process of the novel absolute algorithm, aiming to reduce the dimensions and improve the convergence accuracy. Compared to the improved NR algorithms and the hybrid algorithm, the novel absolute algorithm based on triangular finite elements clustering in the sensitive field has the best effect, which is shown by the experimental results. Keywords：absolute image reconstruction; electrical resistance tomography; inverse problem; real-time; clustering 1　引言 电学层析成像技术包括电磁层析成像 [1-4]、电阻抗层析成像[5-7]、电阻层析成像[8-12]、电容层析成像[13-19]等四种不同分支，而提高图像重建算法求解电阻层析成像逆问题的精度、实时性以及稳定性是