An extreme learning machine combined with Landweber iteration algorithm for the inverse problem of electrical capacitance tomography

作者：Liu Xiao; Wang Xiaoxin; Hu Hongli^*; Li Lin; Yang Xingyue

来源：Flow Measurement and Instrumentation, 2015, 45: 348-356.

DOI：10.1016/j.flowmeasinst.2015.07.009

• Electrical capacitance tomography (ELT)
• Ill-posed
• Extreme learning machine (ELM)
• Landweber iteration

The image reconstruction of the electrical capacitance tomography (ECT) is an ill-posed and sparse problem. In order to increase the accuracy and speed of the image reconstruction, this paper proposes a new reconstruction algorithm which is based on the extreme learning machine (ELM) with the Land-weber iteration method. Firstly, a nonlinear mapping model is established between the pixel gray-scale values and the interelectrode capacitances by using the ELM which has a good learning ability and high speed. Secondly, the Landweber iteration method, which has a good performance in convergence and stability, is applied to calculate the output weight matrix of ELM. Finally, a convergence and stable mapping model of ELM with the Landweber iteration algorithm (L-ELM) for ECT image reconstruction is trained on Matlab platform. Both simulation and measurement tests are carried out to evaluate and analyze the proposed method. Experimental results indicate that the proposed algorithm has good generalization ability and high image reconstruction quality which are better than those of conventional ELM algorithm.

• 出版日期2015-10
• 单位西安交通大学