A Novel Principal Component Analysis Flow Pattern Identification Algorithm for Electrical Capacitance Tomography System

• Published in: 2010 International Conference on Machine Vision and Human-Machine Interface pp. 235 - 238
• Main Authors: Chen Yu, Song Yuchen, Zhang Jian
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.04.2010
• Subjects: Computer networks; Convergence; Data mining; Electrical capacitance tomography; Feature extraction; Fluid flow; Machine vision; neural network; Neural networks; Pattern recognition; Principal component analysis; Symmetric subspace
• ISBN: 9781424465958; 1424465958
• DOI: 10.1109/MVHI.2010.141

To solve the flow pattern identification more difficult problem in electrical capacitance tomography (ECT) technology, a novel principal component analysis flow pattern identification algorithm for neural network is presented. Based on the introduction of the basic principles of feature selection and feature extraction for principal component analysis, Construction of Symmetric subspace model based on principal component analysis neural network, and the convergence of Symmetric subspace algorithm is analyzed. The feasibility of using this algorithm for ECT is also discussed. Algorithm to meet the convergence conditions and to simplify the complex pre-processing steps, greatly reducing the computational complexity, improve the speed of the identification. Experimental results indicate that the algorithm can obtain a higher recognition rate compared with BP neural network recognition algorithm and this new algorithm presents a feasible and effective way to research on flow pattern identification algorithm of electrical capacitance tomography.