Self-adaptive mesh generation in electrical capacitance tomography based on total variation regularization algorithm

Wang Huaxiang,Tang Lei

DOI: https://doi.org/10.3969/j.issn.1002-7300.2006.05.003

2006-01-01

Abstract:To solve ill-posed problem and poor resolution in electrical capacitance tomography, a new image reconstruction algorithm based on total variation regularization is proposed and a new self-adaptive mesh refinement strategy is put forward. Compared with Tikhonov regularization which is commonly used, this new algorithm not only stabilizes the reconstruction, but also enhances the distinguish ability of the reconstruction image in areas with discontinuous medium distribution. It possesses a good edge-preserving property. The self-adaptive mesh generation technique based on this algorithm can refine the mesh automatically in specific areas according to medium distribution. This strategy keeps high resolution as mesh refinement all over the region but reduces calculation loads, thereby speeds up the reconstruction. Experimental results show that this algorithm has advantages in terms of the resolution and real-time performance.