Abstract In order to improve the quality of image reconstruction, in view of the complexity and variety of capacitance data in electrical capacitance tomography (ECT) and the nonlinear relationship between capacitance data and dielectric constant, a deep belief network (DBN) was proposed.The reconstruction algorithm of DBN used the deep nonlinear network structure of DBN to realize the nonlinear relationship between the capacitance value and the gray value of the reconstructed image.The DBN was improved and the adaptive step size (AS) was introduced into the contrastive divergence (CD) algorithm to solve the problem of finding the global optimum with fixed step size and to improve the image quality.In the fine-tuning stage, quasi-Newton method was used to accelerate the convergence speed and reduce the training time.The simulation experiment was carried out on COMSOL 5.3 software, and the image was reconstructed by MATLAB 2014a.The experimental results showed that: DBN can effectively reconstruct the image, and is better than the traditional algorithm; the improved DBN training time is shortened by 5.51s, the image error is as low as 0.0094, and the correlation coefficient is as high as 0.9973, which is a new method and means to study ECT image reconstruction.
［1］曹章. 电学成析成像系统模型、 算法研究［D］. 天津:天津大学, 2007.
［2］李荣瑾. 电学层析成像图像重建研究［D］. 天津:天津大学, 2009.
［3］宋蕾, 陈德运, 姚玉梅, 等. Elman神经网络在ECT系统流型辨识中的应用［J］. 哈尔滨理工大学学报, 2014, 19(5): 103-108.
Song L, Chen D Y, Yao Y M, et al. Application of Elman Neural Network in Pattern Identification for Electrical Capacitance Tomography［J］.Journal of Harbin University of Science and Technology, 2014, 19(5): 103-108.
［4］王莉莉, 刘洪波, 陈德运, 等. 自适应与附加动量BP神经网络的ECT流型辨识［J］. 哈尔滨理工大学学报, 2018, 23(1): 105-110.
Wang L L, Liu H B, Chen D Y, et al. Identification of Flow Regimes Based on Adaptive Learning and Additional Momentum BP Neural Network for Electrical Capacitance Tomography［J］.Journal of Harbin University of Science and Technology, 2018, 23(1): 105-110.
［5］Rosa E D L, Yu W. Randomized algorithms for nonlinear system identification with deep learning modification［J］.Information Sciences, 2016, (364-365): 197-212.
［6］Lecun Y, Bengio Y, Hinton G. Deep learning［J］.Nature, 2015, 521(7553): 436-444.
［7］朱林奇, 张冲, 周雪晴, 等. 融合深度置信网络与与核极限学习机算法的核磁共振测井储层渗透率预测方法［J］.计算机应用, 2017, 37(10): 3034-3038.
Zhu L Q, Zhang C, Zhou X Q, et al. Nuclear magnetic resonance logging reservoir permeability prediction method based on deep belief network and kernel extreme learning machine algorithm［J］.Journal of Computer Applications, 2017, 37(10): 3034-3038.
［8］刘梦溪, 巨永锋, 高炜欣, 等. 焊缝缺陷图像分类识别的深度置信网络研究［J］.测控技术, 2018, 37(8): 5-9.
Liu M X, Ju Y F, Gao W X, et al. Research on Deep Belief Network for Images Classification and Recognition of Weld Defects［J］.Measurement & Control Techlogy, 2018, 37(8): 5-9.
［9］南国芳, 王化祥, 王超. 基于BP神经网络和遗传算法的电阻抗图像重建算法［J］.计量学报, 2003,24(4): 337-340.
Nan G F, Wang H X, Wang C. Image Reconstruction Algorithm for Electrical Impedance Tomography Based on BP Neural Network and Genetic Algorithm［J］.Acta Metrologica Sinica, 2003,24(4): 337-340.
［10］温丽梅, 周苗苗, 李明, 等. 改进的Tikhonov正则化图像重建算法［J］.计量学报, 2018, 39(5): 679-683.
Wen L M, Zhou M M, Li M, et al. Improved Tikhonov Regularization Method for Image Reconstruction［J］.Acta Metrologica Sinica, 2018, 39(05): 679-683.
［11］许鑫. 直接三维电容层析成像传感器优化及重建算法的研究［D］.沈阳:沈阳工业大学, 2014.
［12］王化祥. 电学层析成像技术［J］.自动化仪表, 2017, 38(5): 1-6.
Wang H X. Electrical Tomography Technology［J］.Process Automation Instrumentation, 2017, 38(5): 1-6.
［13］郭琪. 电容层析成像系统的图像重建算法研究［D］.天津:中国民航大学, 2017.
［14］张立峰, 周雷. 基于小波融合的电容层析成像图像重建［J］.计量学报, 2019, 40(2): 285-288.
Zhang L F, Zhou L. Image Reconstruction for Electrical Capacitance Tomography Based on Wavelet Fusion［J］.Acta Metrologica Sinica, 2019, 40(2): 285-288.
［15］马敏,范广永,王涛. 基于ECT的占空比可调双阵列传感器设计［J］.计量学报, 2019, 40(6): 1057-1063.
Ma M, Fan G Y, Wang T. Design of EGR Adjustable Dual Array Sensor Based on ECT［J］.Acta Metrologica Sinica, 2019, 40(6): 1057-1063.
［16］Hinton G E, Osindero S, Teh Y. A fast learning algorithm for deep belief nets［J］.Neural Computation, 2006, 18(7): 1527-1554.
［17］张翔, 石力, 尚勃, 等. 深度置信网络的Spark并行化在微博情感分类中的应用研究［J］.计算机应用与软件, 2018, 35(2): 48-53.
Zhang X, Shi L, Shang B, et al. Research on Spark Parallelization of Deep Confidence Network in Microb-logging Emotion classification［J］.Computer Applicati-ons and Software, 2018, 35(2): 48-53.
［18］时培明, 赵娜, 苏冠华, 等. 变载荷齿轮箱故障信号智能检测方法［J］.计量学报, 2018, 39(6): 847-851.
Shi P M, Zhao N, Su G H, et al. Intelligent Detection Method of Variable Load Gearbox Fault Signal［J］.Acta Metrologica Sinica, 2018, 39(6): 847-851.
［19］Fischer A, Igel C. Training restricted Boltzmann machines: An introduction［J］.Pattern Recognition, 2014, 47(1): 25-39.
［20］周树森. 基于深度置信网络的分类方法［M］.北京:清华出版社, 2015.
［21］Lopes N, Ribeiro B. Towards adaptive learning with improved convergence of deep belief networks on graphics processing units［J］.Pattern Recognition, 2014, 47(1): 114-127.
［22］Wu G Z, Liang H Y. A modified BFGS method and its convergence［J］.Computer Modelling & New Technologies,2014, 18(11): 43-47.
2021, Vol. 42 
