Data preprocessing methods for electrical impedance tomography: A review

Objective: Electrical impedance tomography (EIT) is a promising measurement technique in applications, especially in industrial monitoring and clinical diagnosis. However, two major drawbacks exist that limit the spatial resolution of reconstructed EIT images, i.e. the 'soft field' effect and the ill-posed problem. In recent years, apart from the development of reconstruction algorithms, some preprocessing methods for measured data or sensitivity maps have also been proposed to reduce these negative effects. It is necessary to find the optimal preprocessing method for various EIT reconstruction algorithms. Approach: In this paper, seven typical data preprocessing methods for EIT are reviewed. The image qualities obtained using these methods are evaluated and compared in simulations, and their applicable ranges and combination effects are summarized. Main results: The results show that all the reviewed methods can enhance the quality of EIT reconstructed images to different extents, and there is an optimal one under any given reconstruction algorithm. In addition, most of the reviewed methods do not work well when using the Tikhonov regularization algorithm. Significance: This paper introduces the preprocessing method to EIT, and the quality of reconstructed images obtained using these methods is evaluated through simulations. The results can provide a reference for practical applications. © 2020 Institute of Physics and Engineering in Medicine