Feature Driven Algebraic Reconstruction Technique Algorithm for Electrical Tomography

Electrical resistance tomography (ERT) is an advanced visualization detection technology with non-invasive, non-radiative and fast-response advantages. At present, it has become an important technique in industrial process detection. Image reconstruction algorithm is one of the research focuses, among which algebraic reconstruction technique (ART) has widely used in tomography process. However, the existing ART algorithm is subject to slow convergence, low spatial resolution and poor interpretability in the ERT reconstruction process. To overcome these problems, this paper presents a feature-driven ART reconstruction algorithm. This algorithm takes each group of measurements from each excitation electrode as a feature to optimize the convergence process of the traditional ART algorithm. Both simulations and experiment results show that the new ART algorithm and its variants are better than the original one in terms of both temporal resolution and spatial resolution.