|
|
|
| Image Reconstruction Algorithm for Electrical Capacitance Tomography Based on BP Neural Network |
| MA Min1,GAO Zhen-fu1,WANG Hua-xiang2 |
1.Civil Aviation University of China, Tianjin 300300, China;
2.School of Electrical Engineering & Automation, Tianjin University, Tianjin 300072, China |
|
|
|
|
Abstract Considering the electrical capacitance tomography image reconstruction is an illposed problem, the system model is set up in COMSOL software, resolving the Forward problem. BP neural network is applied to realize the nonlinear mapping between the capacitance values and the reconstructed image grey values based on the nonlinear mapping and the association memory ability. Compared with the traditional ones, the algorithm avoids the trivial solution for the sensitivity matrix, and the linearization that reduces the imaging accuracy. Two filtering methods are proposed to improve image quality, and the simulation was completed in the MATLAB workbench.
|
|
|
|
|
|
［1］彭黎辉,陆耿,杨五强. 电容成像图像重建算法原理及评价［J］. 清华大学学报, 2004, 44(3): 472-475.
［2］雷兢. 多相流的电容层析成像图像重建研究［D］. 北京:中国科学院研究生院（工程热物理研究所）, 2008.
［3］李久贤,孙伟,夏良正. 一种新的模糊对比度增强算法 ［J］. 东南大学学报,2004,34(5) :675-677.
［4］王化祥,杨五强. 电容过程成像技术的进展［J］. 仪器仪表学报, 2000,21(1): 4-6.
［5］何世钧. 电容层析成像系统的应用与研究［D］ . 天津: 天津大学,2005: 9-14.
［6］孙强. 电容层析成像油水两相流测量系统开发平台设计［D］. 东营: 中国石油大学(华东),2009,23-27.
［7］王化祥,张立峰. 电学层析成像激励测量模式及图像重建算法的研究［D］. 天津: 天津大学,2010,51-58.
［8］韩力群. 人工神经网络理论设计及应用［M］ . 北京: 化学工业出版社,2007,61-79.
［9］丛爽. 面向MATLAB工具箱的神经网络理论与应用［M］. 合肥: 中国科学技术大学出版社,2009,81-133.
［10］Ortiz-Aleman C,Martin R.Two-phase oil-gas pipe flow imaging by simulated annealing［J］. Geophys Eng,2005,2: 32-37.
［11］马平,周晓宇,田沛,等. 基于COMSOL电容层析成像［J］. 电测与仪表. 2009,(11): 20-23.
［12］靳蕃. 神经计算智能基础［M］. 成都: 西南交通大学出版社,2000
［13］Reed R. Pruning algorithm—a survey［J］. IEEE Trans on Neural Networks,1993,4(5): 740-747.
［14］张德丰.MATLAB神经网络编程［M］. 北京:化学工业出版社,2011,131-150.
［15］张德贤. 前向神经网络合理隐含层结点个数估计［D］. 郑州: 郑州工程学院,2003:21-23.
［16］张汗灵. MATLAB在图像处理中的应用［M］. 北京: 清华大学出版社,2011,71-100. |
|
|
|
2013, Vol. 34 
