Research on a Method to Eliminate Redundant Data in Electrical Capacitance Tomography
Hang Long, Lina Ma, Nan Li
Abstract
To reduce the influence of redundant data on imaging accuracy in the process of electrical capacitance tomography (ECT) image reconstruction, we proposed a method to eliminate redundant data in capacitance tomography based on the hierarchical clustering (HC) algorithm. According to the characteristics of the capacitance data obtained by the sensor, we used an HC algorithm to classify the capacitance data. Through the simulation research of four typical medium distributions, we arrived at the preliminary conclusion that there were more redundant data when the medium was close to the sensor electrode, whereas the data redundancy phenomenon was not obvious when the medium was located in the center of the sensor pipeline. Simulations and experiments were carried out for the medium distributions of several mediums close to the electrodes. The results showed that the quality of the reconstructed images obtained by using the HC algorithm to eliminate redundant data was significantly better than the reconstructed images obtained by using all the data. For the laminar flow reconstructed images, the relative image errors (REs) of the simulation and the experiment were reduced by at most 0.1683 and 0.1606, and the correlation coefficients (CCs) were increased by at most 0.1012 and 0.0679. For the reconstructed images of the annular flow, the REs of the simulation and experiment were reduced by at most 0.1519 and 0.0598, and the CCs were increased by at most 0.0765 and 0.0455.