Research on Parallel Policy of Image Reconstruction of Electrical Capacitance Tomography System

Image reconstruction is a key to electrical capacitance tomography system. In this paper, basic principle and image construction problem of ECT system is analyzed, and BP neural network is adopted to solve image reconstruction. To accelerate reconstruction, BP neural network is divided into several parts, and each part will be sent to different slave nodes on which Globus platform is installed. After every part is computed in slave nodes, the results will be sent to the master node and a frame of reconstruction image is reorganized. The results show that this method adopted is feasible, and it has the advantages of high precise, speediness.